On Thu, May 19, 2011 at 7:35 PM, Benjamin Herrenschmidt
[off-list ref] wrote:

On Wed, 2011-05-18 at 16:24 -0500, Eric Van Hensbergen wrote:

quoted

+ =A0 =A0 =A0 =A0 =A0 =A0 .dcache_bsize =A0 =A0 =A0 =A0 =A0 =3D 32,
+ =A0 =A0 =A0 =A0 =A0 =A0 .cpu_setup =A0 =A0 =A0 =A0 =A0 =A0 =A0=3D __se=

tup_cpu_460gt,

=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =

=A0 =A0 =A0 =A0^^^^^^^^^^^^^^^^^^

Are you sure ?

That surprised me too, I figured it must have been a close enough
match (at least in the 2.6.29.1 time frame which is where I'm trying
to merge the BG/P patches up from.  The kittyhawk patches don't even
use this, so its possible we could just remove it.


_GLOBAL(__setup_cpu_460ex)
_GLOBAL(__setup_cpu_460gt)
        mflr    r4
        bl      __init_fpu_44x
        bl      __fixup_440A_mcheck
        mtlr    r4
        blr

Looks like the 460 setup invokes a bunch of 440 calls!  Would you
prefer I setup my own entry point (setup_cpu_bgp or setup_cpu_450)
which makes the same calls?

     -eric