In article <Pine.SOL.4.33.0107291649400.16494-100000@draco.cus.cam.ac.uk>,
Ben Harris <bjh21@netbsd.org> wrote:
>
>loadfile, though, loads the section headers, and then all the STRTAB and
>SYMTAB sections, which means it has to seek backwards to get at .shstrtab.
>There's no need for it to load .shstrtab, though, since this just provides
>names for the sections, not for any symbols.
>
>What this patch does is to arrange that before deciding to load a STRTAB
>section, loadfile checks if any SYMTAB section references it.  If none
>does, it skips the section.  This seems to prevent loadfile having to seek
>backwards at all when loading arm26 kernels, which seems good to me.
>
>Could someone responsible for loadfile take a look at this and tell me
>what they think?
>

Looks ok, (except for a mis-indentation of "first = 0;"). If the kernel
debugger works with the new loadfile we should try it out.

christos