[oe] Puzzled over Angstrom generated kernel on a new board

[Daniel P. Smith]

> -----Original Message-----
> From: openembedded-devel-bounces at lists.openembedded.org
> [mailto:openembedded-devel-bounces at lists.openembedded.org] On Behalf Of
> Koen Kooi
> Sent: Tuesday, May 10, 2011 2:05 PM
> To: openembedded-devel at lists.openembedded.org
> Subject: Re: [oe] Puzzled over Angstrom generated kernel on a new board
> 
> 
> Try defining
> 
> UBOOT_ENTRYPOINT = "0x80008000"
> UBOOT_LOADADDRESS = "0x80008000"
> 
> You probably need to change the values to match your hw, though

Haha, my apologies and thanks as I was looking through the bitbake logs for the kernel and found that the recipe recreates the uImage file setting the entry and load addr's to 0x20008000, where as the CNS3xxx is expecting 0x00008000. But this brings up the question for me is why does the kernel recipes recreate the uImage? The make uImage command will/should pull the zreladdr from the sub-arch's Makefile.boot. I would assume it is so OE can provide the flexibility, but I don't like to make assumptions.


On a side note, has anyone considered cataloging the variety of variables that can be set by a recipe author? I would have never known about the two Koen pointed out had I not asked here or spent hours poring over the internals to find out where 0x20008000 was coming from.


V/r,
Daniel P. Smith