readahead: fault retry breaks mmap file read random detection
commit45cac65b0fcd287ebb877b141d40ba9bbe8e5da7
authorShaohua Li <shli@kernel.org>
Mon, 8 Oct 2012 23:32:19 +0000 (8 16:32 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 9 Oct 2012 07:22:47 +0000 (9 16:22 +0900)
tree30ed25c91aaeed153de51a78d171cb14582e383f
parente79bee24fd6134f90af4228cfebd010136d67631
readahead: fault retry breaks mmap file read random detection

.fault now can retry.  The retry can break state machine of .fault.  In
filemap_fault, if page is miss, ra->mmap_miss is increased.  In the second
try, since the page is in page cache now, ra->mmap_miss is decreased.  And
these are done in one fault, so we can't detect random mmap file access.

Add a new flag to indicate .fault is tried once.  In the second try, skip
ra->mmap_miss decreasing.  The filemap_fault state machine is ok with it.

I only tested x86, didn't test other archs, but looks the change for other
archs is obvious, but who knows :)

Signed-off-by: Shaohua Li <shaohua.li@fusionio.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
20 files changed:
arch/arm/mm/fault.c
arch/avr32/mm/fault.c
arch/cris/mm/fault.c
arch/hexagon/mm/vm_fault.c
arch/ia64/mm/fault.c
arch/m68k/mm/fault.c
arch/microblaze/mm/fault.c
arch/mips/mm/fault.c
arch/openrisc/mm/fault.c
arch/powerpc/mm/fault.c
arch/s390/mm/fault.c
arch/sh/mm/fault.c
arch/sparc/mm/fault_32.c
arch/sparc/mm/fault_64.c
arch/tile/mm/fault.c
arch/um/kernel/trap.c
arch/x86/mm/fault.c
arch/xtensa/mm/fault.c
include/linux/mm.h
mm/filemap.c