On Mon, 2009-12-07 at 15:42 -0600, Scott Wood wrote: > Wolfgang Denk wrote: > > Dear Peter Tyser, > > > > In message <1258943983.9357.20.ca...@ptyser-laptop> you wrote: > >> My personal preference would be to take it a step further and organize > >> like: > >> > >> /arch/$ARCH/ > >> /lib/<sources currently in lib_$ARCH/> > >> /include/asm/<headers currently in incluce/asm-$ARCH/> > >> /cpu/<appropriate cpu directories in cpu/> > >> /config.mk (old lib_$ARCH/config.mk) > >> > >> /lib/ > >> /<source files currently in lib_generic> > >> /libfdt/ > >> /lzma/ > >> /lzo/ > >> > >> This layout would more closely match the Linux layout and is cleaner > >> overall in my opinion. What do others think of this idea? > > > > I generally agree with the idea, with the exception of > > /include/asm/<headers> - I really hate this change in the Linux > > kernel, and still fail to see any advantages of this. > > I really like only having one subtree to grep for arch stuff, and it > would avoid presenting files that are in git, under a path that git > doesn't know about. > > Plus, the symlink is one less thing that the makefiles would need to > keep in sync when arch changes.
I agree with Scott, especially the point about grepping for arch stuff in one directory. It also seems like a generally logical separation of arch-specific code/headers. What is the downside of splitting out the headers to arch/ directories? Or the advantage of keeping them in include/asm-<arch>? Best, Peter _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot