i've done a dozen random tests too and it's looking good so far. Nice
work!
pgtable_32.h and pgtable_64.h still look a tiny bit messy from the
include file dependencies POV. For example pgtable_32.h:
#include <asm/processor.h>
#include <asm/fixmap.h>
#include <linux/threads.h>
#include <asm/paravirt.h>
#include <linux/bitops.h>
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/spinlock.h>
that asm/paravirt.h include is already present in pgtable.h, in a
somewhat quirky way:
#ifdef CONFIG_PARAVIRT
#include <asm/paravirt.h>
#else /* !CONFIG_PARAVIRT */
also, most of the:
scripts/checkpatch.pl --file include/asm-x86/pgtable*.h
complaints are real ones and should be fixed.
would you be interested in cleaning up that stuff once and forever? It
would be a fine approach if you just tried to quickly push for a "high
quality" end result in a series of patches and sent that series to me,
without having tested it fully through - i can figure out whatever build
breakages and dependencies there still are. So there would be no
expectation of getting such a cleanup series right in the first (or
second, or third) attempt, this is spaghetti code that has been
accumulated up for years. The important thing would be to be careful to
not introduce runtime breakages accidentally - build breakages due to
some include file dependency we can sort out just fine. Hm?
Ingo
--
