On 03/15/2013 10:22:06 AM, Johannes Berg wrote:

On Thu, 2013-03-14 at 11:52 -0500, Scott Wood wrote:

quoted

quoted

quoted

+#ifdef CONFIG_PPC_BOOK3S_64
 	/* can't use RESTORE_SPECIAL(MSR) */
 	ld	r0, SL_MSR(r11)
 	mtmsrd	r0, 0

Unfortunately, I forgot the reason for this comment, and didn't =20

put a

quoted

quoted

better one (almost 6 years ago!!)

=20

quoted

If it's because book3s needs mtmsrd instead of mtmsr, that doesn't
apply to booke.

=20
Indeed, looking at the code again now that seems pretty obvious.
=20
Looking at the patch again, I'd be a little concerned about the lack =20
of
cache flushing, seems a bit odd but I'm sure you know what you're =20
doing
(and I don't know book3e at all, and hardly remember book3s -- or even
that name...) :-)

Could you elaborate on why book3s flushes the way it does?  What's =20
special about the first 32 MiB?  If it's to cover kernel code, why =20
would that be changing from what's already there?

-Scott=