On Fri, Apr 17, 2009 at 03:06, Michael Schmitz <[email protected]> wrote: >> > Geert, Roman: does paging_init rely on part of the first memory chunk being >> > mapped in head.S, or just part of the chunk the kernel resides in? Could I >> > reorder the m68k_memory[] elements in setup.c to fix the ordering problems? >> >> Should work. config_amiga() removes Zorro II memory chunks, and that's >> just before >> paging_init() is called. of course you cannot reorder the first chunk, >> as that one is >> already in use. > > Found that out the hard way. So what's the fix for this? Does the first chunk > always have to be the one the kernel resides in? > > What is the reason for dropping chunks that are residing below the fisrst one, > in paging_init ??
`git blame' tells you that code was added by commit 12d810c1b8c2b913d48e629e2b5c01d105029839 Author: Roman Zippel <[email protected]> Date: Thu May 31 00:40:54 2007 -0700 m68k: discontinuous memory support Probably discontigmem supports only increasing physical addresses of the multiple memory chunks :-( But the newer sparsemem generic model may support it, see Documentation/memory-hotplug.txt. Gr{oetje,eeting}s, Geert -- Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected] In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that. -- Linus Torvalds -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
