This patch adds support for OMAP3 platforms. Mainly to setup the infrastructure.
ARMV7 requires a different I/D cache cleanup code which is introduced in this patch
Signed-off-by: Nishanth Menon<x0nishan@ti.com>
Small comment: the c7,c7,0 operation for v4 above will invalidate
both I and D caches whereas the Cortex-A8 code will only
invalidate I cache. (BTW the comment for v4 should say
"invalidate" instead of "flush" for c7,c7,0.)
The code following these invalidations explicitly disables D cache,
but if some code later on switches on the D cache without any
D cache invalidation, bad things will happen 
Laurent
From: beagleboard@googlegroups.com [mailto:beagleboard@googlegroups.com] On Behalf Of Laurent
Desnogues
Sent: Wednesday, May 21, 2008 2:04 AM
To: beagleboard@googlegroups.com
Subject: [beagleboard] Re: [RFC][Patch 1/4] U-Boot-V2:ARM:OMAP3: Add support for OMAP and Cortex A8Small comment: the c7,c7,0 operation for v4 above will invalidate
both I and D caches whereas the Cortex-A8 code will only
invalidate I cache.
Yep. I have revamped the patch as follows (code from OMAP-UbootV1 cpu/omap3/start.S - I have skipped the smi part of things. It is a bit weird. In EMU/HS devices, I need to call PPA functions to handle it. So have provided a Kconfig option to handle it.
(BTW the comment for v4 should say
"invalidate" instead of "flush" for c7,c7,0.)
Yep. but that needs to go in as a separate patch. Could you post one to the U-Boot users? Else will do so 2morrow. Kinda
The code following these invalidations explicitly disables D cache,
but if some code later on switches on the D cache without any
D cache invalidation, bad things will happen
Hmmm...
Signed-off-by: Nishanth Menon<x0nishan@ti.com>