Re: [Cbe-oss-dev] [RFC 3/3] powerpc: copy_4K_page tweaked for Cell

[Mark Nelson <markn@au1.ibm.com>]

On Fri, 20 Jun 2008 07:28:50 am Arnd Bergmann wrote:
> On Thursday 19 June 2008, Mark Nelson wrote:
> > =A0=A0=A0=A0=A0=A0=A0=A0.align =A07
> > _GLOBAL(copy_4K_page)
> > =A0=A0=A0=A0=A0=A0=A0=A0dcbt=A0=A0=A0=A00,r4=A0=A0=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0/* Prefetch ONE SRC cacheline */
> >=20
> > =A0=A0=A0=A0=A0=A0=A0=A0addi=A0=A0=A0=A0r6,r3,-8=A0=A0=A0=A0=A0=A0=A0=
=A0/* prepare for stdu */
> > =A0=A0=A0=A0=A0=A0=A0=A0addi=A0=A0=A0=A0r4,r4,-8=A0=A0=A0=A0=A0=A0=A0=
=A0/* prepare for ldu */
> >=20
> > =A0=A0=A0=A0=A0=A0=A0=A0li=A0=A0=A0=A0=A0=A0r10,32=A0=A0=A0=A0=A0=A0=A0=
=A0=A0=A0/* copy 32 cache lines for a 4K page */
> > =A0=A0=A0=A0=A0=A0=A0=A0li=A0=A0=A0=A0=A0=A0r12,128+8=A0=A0=A0=A0=A0=A0=
=A0=A0=A0=A0=A0=A0=A0=A0=A0/* prefetch distance*/
>=20
> Since you have a loop here anyway instead of the fully unrolled
> code, why not provide a copy_64K_page function as well, jumping in
> here?

That is a good idea. What effect will that have on how the code
patching will work?

>=20
> The inline 64k copy_page function otherwise just adds code size,
> as well as being a tiny bit slower. It may even be good to
> have an out-of-line copy_64K_page for the regular code, just
> calling copy_4K_page repeatedly.

Doing that sounds like it'll make the code patching easier.

Thanks!

Mark