On Tue, Dec 13, 2011 at 11:18 AM, Wolfgang Denk wd@denx.de wrote:
Dear Tom,
In message CA+M6bX=j+=UBkZ-VGoAJgFtfap+sLp4i2JypDSjJ_z-bCbw6sQ@mail.gmail.com you wrote:
Which covers the omap config file changes. Assuming Wolfgang doesn't see a problem with using SYS_... in defines, Scott, do you want this via your tree (in /next) since it's NAND or in mine since it's TI boards? Assuming you don't see any problems of course. Thanks!
Indeed.
Stefano, what is this SYS_NAND_* supposed to mean? Either it is a config option, then you should name it CONFIG_SYS_NAND_* (or CONFIG_NAND_), or it is something else, in which case the "SYS_" is difficult to swallow for me (unless you have a really good explanation?)
So, we have various characteristics of the NAND chip that need to be described for SPL. Today we do CONFIG_SYS_NAND... but they're unchanged in every implementation and are calculated from other per-board values. So to avoid duplication in each config file, we drop them from there and use them only locally #define calculate them in the file that needs them. Since they're still system specific values, I had suggested calling them SYS_... but I wasn't sure what the right namespace is. I had a recollection that using CONFIG_ outside of config files was a bad idea.