[U-Boot] [PATCH 10/10] board/t104xrdb: Add support of NAND, SD, SPI boot for T1040RDB

[Rommel G Custodio]

Dear Prabhakar Kushwaha,

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

> 
> 
> On 4/1/2014 4:40 AM, Rommel G Custodio wrote:
> > Dear Prabhakar Kushwaha,
> >
> > Prabhakar Kushwaha <prabhakar <at> freescale.com> writes:
> >


8>< snipped ><8

> >> +#Flush PBL data
> >> +091380c0 000FFFFF
> > That's a "Wait" command.
> > Unless the "Flush" was command (that is in most PBI files I've read,
> > Flush/Wait pait) was inadvertantly deleted.
> >
> 
> Flush command only work for CCSRBAR address space. It should not be used 
> for CPC-SRAM.
> 
> This is the reason, it has been removed.

Does this affect all mpc85xx cores?

I've noticed this same weird behavior on a T1040QDS.
A simple write to CPC-SRAM before a Flush command causes the LEDs to report 
an error. 
i.e
89ffff00 4bfff004
09138000 00000000
091380c0 00000001

After putting the CPC-SRAM write after the Flush the write was successful 
(as verified by a BDI-3000).
i.e
09138000 00000000
091380c0 00000001
89ffff00 4bfff004
091380c0 00000001


BUT it seems this does not occur on the P5040DS. The P5040DS_SPIFLASH 
creates a u-boot that has Flush commands and it boots correctly.
[local]$ xxd -g4 obj-P5040DS_SPIFLASH/u-boot.pbl
:
00cc0a0: ffffffff 81ffffc0 ffffffff ffffffff  ................
00cc0b0: ffffffff ffffffff ffffffff ffffffff  ................
00cc0c0: ffffffff ffffffff ffffffff ffffffff  ................
00cc0d0: ffffffff ffffffff ffffffff ffffffff  ................
00cc0e0: ffffffff 4bfff004 09138000 00000000  ....K...........
00cc0f0: 091380c0 00000000 08138040 aa69eb89  ........... at .i..



On another note, if this affects all mpc85xx cores then I guess 
tools/pblimage.c needs updating. The add_end_cmd() always adds a Flush 
command prior to a CRC check (CONT=0) command.

> 
> Regards,
> Prabhakar
> 

All the best,
Rommel