On 08/09/2012 08:03 AM, Kirill A. Shutemov wrote:

From: Andi Kleen <redacted>

Add a cache avoiding version of clear_page. Straight forward integer variant
of the existing 64bit clear_page, for both 32bit and 64bit.

Also add the necessary glue for highmem including a layer that non cache
coherent architectures that use the virtual address for flushing can
hook in. This is not needed on x86 of course.

If an architecture wants to provide cache avoiding version of clear_page
it should to define ARCH_HAS_USER_NOCACHE to 1 and implement
clear_page_nocache() and clear_user_highpage_nocache().

Compile failure:

/home/hpa/kernel/tip.x86-mm/arch/x86/mm/fault.c: In function 
‘clear_user_highpage_nocache’:
/home/hpa/kernel/tip.x86-mm/arch/x86/mm/fault.c:1215:30: error: 
‘KM_USER0’ undeclared (first use in this function)
/home/hpa/kernel/tip.x86-mm/arch/x86/mm/fault.c:1215:30: note: each 
undeclared identifier is reported only once for each function it appears in
/home/hpa/kernel/tip.x86-mm/arch/x86/mm/fault.c:1215:2: error: too many 
arguments to function ‘kmap_atomic’
In file included from 
/home/hpa/kernel/tip.x86-mm/include/linux/pagemap.h:10:0,
                  from 
/home/hpa/kernel/tip.x86-mm/include/linux/mempolicy.h:70,
                  from 
/home/hpa/kernel/tip.x86-mm/include/linux/hugetlb.h:15,
                  from /home/hpa/kernel/tip.x86-mm/arch/x86/mm/fault.c:14:
/home/hpa/kernel/tip.x86-mm/include/linux/highmem.h:66:21: note: 
declared here
make[4]: *** [arch/x86/mm/fault.o] Error 1
make[3]: *** [arch/x86/mm] Error 2
make[2]: *** [arch/x86] Error 2
make[1]: *** [sub-make] Error 2
make[1]: Leaving directory `/home/hpa/kernel/tip.x86-mm'

This happens on *all* my test configurations, including both x86-64 and 
i386 allyesconfig.  I suspect your patchset base is stale.

	-hpa

-- 
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.