On 4/30/07, David H. Lynch Jr. <[EMAIL PROTECTED]> wrote: > There are a plethora of bd_info's in the kernel tree. I think that > u-boot is probably the most common, > but the u-boot one is huge and full of cruft.
Be realistic; we're talking about 124 bytes (vs 64 bytes for the pico bd_t). If memory consumption is a concern, there are far more interesting targets. ppcboot.h is butt-ugly, but at least it's somewhat common. > There are also likely to be any number for board/loader combinations > that are not in the kernel. If its not in mainline, then I'm not concerned about it because then you're patching anyway. I made sure it is easy to patch in a new bd_t so nobody is screwed. When those board ports are pushed for mainline inclusion, then we can add the appropriate conditionals to virtex.h, but in the mean time it's a moot point. > But I would prefer that if there is a bd_info struct that it be defined > by the board not virtex.h or virtex.c. Understood, but I do not want to encourage that approach between now and when we jump to arch/powerpc (legacy bootloader support, yadda, yadda). Cheers, g. -- Grant Likely, B.Sc. P.Eng. Secret Lab Technologies Ltd. [EMAIL PROTECTED] (403) 399-0195 _______________________________________________ Linuxppc-embedded mailing list [email protected] https://ozlabs.org/mailman/listinfo/linuxppc-embedded
