On Jan 31, 2011, at 3:37 PM, Wolfgang Denk wrote:
Dear Haiying Wang,
In message 1296507737.2049.518.camel@haiying-laptop you wrote:
On Mon, 2011-01-31 at 21:08 +0100, Wolfgang Denk wrote:
Dear Haiying.Wang@freescale.com,
In message 1296499317-26616-6-git-send-email-Haiying.Wang@freescale.com you wrote:
From: Haiying Wang Haiying.Wang@freescale.com
For some board which doesn't have NOR flash and the QE's firmware(ucode) is saved in its NAND flash, we don't want call qe_init in cpu_init_r, but will call it later after nand is initialized.
Is there a pressing reason to do this so early for other boards? Can not all boards initialize this later?
My understanding is that QE is a cpu feature, so it is called early in cpu_init_r. As Kumar recommended before, I can move qe_init from cpu_init_r to misc_init_r for every 85xx boards with qe support. Is it acceptable to you?
Yes, if this way we can avoid to do the same thing at different points in the initialization sequence.
Doing this in misc_init_r() isn't right either.
We've had this argument before, can we just add a cpu_init_late_r() that is post env_relocate() ?
Why should we duplicate cpu generic code in board code?
- k