On Thursday 22 May 2008, Grant Erickson wrote:
Why do we now still have the SDRAM setup for Kilauea included in assembler? Is it because of the NAND booting support? If yes, then leave it as is and I'll clean this up later.
Precisely. You have better exposure and visibility to the subtle nuances of NAND booting on Kilauea, so wrapping up that particular issue is best left to your experience.
OK. I'll take care of this.
diff --git a/board/amcc/kilauea/memory.c b/board/amcc/kilauea/memory.c index 1d7a3fa..b7e2344 100644 --- a/board/amcc/kilauea/memory.c +++ b/board/amcc/kilauea/memory.c
...
diff --git a/board/amcc/makalu/memory.c b/board/amcc/makalu/memory.c index b03b60b..326feb7 100644 --- a/board/amcc/makalu/memory.c +++ b/board/amcc/makalu/memory.c
...
Please remove this file completely. It's not needed anymore, since we now use the common DDR2 controller init code you introduced with this patch. Please remove the testdram() function completely and move the sdram_init() to makalu.c.
For both Kilauea and Makalu or just Makalu and you'll address Kilauea in the course of handling init.S?
Just do Makalu. I need to tackle Kilauea anyway, so I'll clean this up in a later patch in this merge window.
And please base you new patches against the TOT of Wolfgangs master branch. He has already pushed his "Big white-space cleanup" and there could be merge problems with your work.
Thanks.
Best regards, Stefan
===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office@denx.de =====================================================================