[U-Boot] [RFC 5/5] B4860QDS: Add support of 2 stage NAND boot loader

Rommel Custodio sessyargc+uboot at gmail.com
Fri Sep 20 01:21:41 CEST 2013


Dear Prabhakar Kushwaha,

Prabhakar Kushwaha <prabhakar <at> freescale.com> writes:
> > 
> > 

<snipped>

Thank you for the replies.

> > You mention you use the PBL... but probably not a pblimage. The patch
> > correctly fixes RAMBOOT_PBL as a trigger to generate the pblimage (u-
> > boot.pbl) but there seems to be no RCW or PBI file defined.
> 
> B4860 does support PBL based NAND boot.
> but I have yet to integrate u-boot-spl.bin to generated u-boot-spl.pbl.
> This is in my TODO list

This is my understanding (if need be, take it with a grain of salt)

The Makefiles need to be modified so that SPL build will create a pblimage.
The "mkimage pblimage" needs to be executed only on the SPL binary. Then the  
top-level Makefile proceeds in concatenating the u-boot-spl.bin (now a 
pblimage) and u-boot.bin.


> > I have one question, can this scenario be implemented on a P5040? i.e.
> > simulate that CPC is around 128Kb and load u-boot via SPL? Now all 
corenet
> > processors seem to only support pblimage booting.
> >
> 
> yes. This scenario can be implemented for P5040. Only need to create 
> spl.c file + define constants.

Seems easy enough :-) though I probably don't have the resources to actually 
implement it.

Just a note. The pblimage booting is not very flexible now. Currently it 
assumes that u-boot.pbl will fit into CPC (configured as SRAM). This is OK 
for most processors (i.e P5040 with 1Mb CPC) but it will not work for others 
(i.e T1040/1042 with only 256Kb CPC). So there is a need to have SPL in this 
case.

All the best,
Rommel



More information about the U-Boot mailing list