linux.git/drivers/staging/zsmalloc, branch v3.10

staging/zsmalloc: don't use pgtable-mapping from modules

2013-04-23T17:34:20+00:00

Building zsmalloc as a module does not work on ARM because it uses
an interface that is not exported:

ERROR: "flush_tlb_kernel_range" [drivers/staging/zsmalloc/zsmalloc.ko] undefined!

Since this is only used as a performance optimization and only on ARM,
we can avoid the problem simply by not using that optimization when
building zsmalloc it is a loadable module.

flush_tlb_kernel_range is often an inline function, but out of the
architectures that use an extern function, only powerpc exports
it.

Signed-off-by: Arnd Bergmann 
Cc: Greg Kroah-Hartman 
Cc: Seth Jennings 
Cc: Nitin Gupta 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: Fix link error on ARM

2013-03-28T23:08:54+00:00

Testing the arm chromebook config against the upstream
kernel produces a linker error for the zsmalloc module from
staging. The symbol flush_tlb_kernel_range is not available
there. Fix this by removing the reimplementation of
unmap_kernel_range in the zsmalloc module and using the
function directly. The unmap_kernel_range function is not
usable by modules, so also disallow building the driver as a
module for now.

Cc: stable 
Signed-off-by: Joerg Roedel 
Acked-by: Minchan Kim 
Signed-off-by: Greg Kroah-Hartman

mm: rename page struct field helpers

2013-02-24T01:50:18+00:00

The function names page_xchg_last_nid(), page_last_nid() and
reset_page_last_nid() were judged to be inconsistent so rename them to a
struct_field_op style pattern.  As it looked jarring to have
reset_page_mapcount() and page_nid_reset_last() beside each other in
memmap_init_zone(), this patch also renames reset_page_mapcount() to
page_mapcount_reset().  There are others like init_page_count() but as
it is used throughout the arch code a rename would likely cause more
conflicts than it is worth.

[akpm@linux-foundation.org: fix zcache]
Signed-off-by: Mel Gorman 
Suggested-by: Andrew Morton 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

staging: zsmalloc: remove unused pool name

2013-01-30T17:22:41+00:00

zs_create_pool() currently takes a name argument which is
never used in any useful way.

This patch removes it.

Signed-off-by: Seth Jennings 
Acked-by: Nitin Gupta 
Acked-by: Rik van Riel 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: Fix TLB coherency and build problem

2013-01-30T04:22:16+00:00

Recently, Matt Sealey reported he fail to build zsmalloc caused by
using of local_flush_tlb_kernel_range which are architecture dependent
function so !CONFIG_SMP in ARM couldn't implement it so it ends up
build error following as.

  MODPOST 216 modules
  LZMA    arch/arm/boot/compressed/piggy.lzma
  AS      arch/arm/boot/compressed/lib1funcs.o
ERROR: "v7wbi_flush_kern_tlb_range"
[drivers/staging/zsmalloc/zsmalloc.ko] undefined!
make[1]: *** [__modpost] Error 1
make: *** [modules] Error 2
make: *** Waiting for unfinished jobs....

The reason we used that function is copy method by [1]
was really slow in ARM but at that time.

More severe problem is ARM can prefetch speculatively on other CPUs
so under us, other TLBs can have an entry only if we do flush local
CPU. Russell King pointed that. Thanks!
We don't have many choices except using flush_tlb_kernel_range.

My experiment in ARMv7 processor 4 core didn't make any difference with
zsmapbench[2] between local_flush_tlb_kernel_range and flush_tlb_kernel_range
but still page-table based is much better than copy-based.

* bigger is better.

1. local_flush_tlb_kernel_range: 3918795 mappings
2. flush_tlb_kernel_range : 3989538 mappings
3. copy-based: 635158 mappings

This patch replace local_flush_tlb_kernel_range with
flush_tlb_kernel_range which are avaialbe in all architectures
because we already have used it in vmalloc allocator which are
generic one so build problem should go away and performane loss
shoud be void.

[1] f553646, zsmalloc: add page table mapping method
[2] https://github.com/spartacus06/zsmapbench

Cc: stable@vger.kernel.org
Cc: Dan Magenheimer 
Cc: Russell King 
Cc: Konrad Rzeszutek Wilk 
Cc: Nitin Gupta 
Cc: Seth Jennings 
Reported-by: Matt Sealey 
Signed-off-by: Minchan Kim 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: make CLASS_DELTA relative to PAGE_SIZE

2013-01-30T04:16:42+00:00

Right now ZS_SIZE_CLASS_DELTA is hardcoded to be 16.  This
creates 254 classes for systems with 4k pages. However, on
PPC64 with 64k pages, it creates 4095 classes which is far
too many.

This patch makes ZS_SIZE_CLASS_DELTA relative to PAGE_SIZE
so that regardless of the page size, there will be the same
number of classes.

Acked-by: Nitin Gupta 
Acked-by: Minchan Kim 
Signed-off-by: Seth Jennings 
Acked-by: Dan Magenheimer 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: comment zs_create_pool function

2013-01-16T07:49:55+00:00

Just as with zs_malloc() and zs_map_object(), it is worth
formally commenting the zs_create_pool() function.

Signed-off-by: Davidlohr Bueso 
Signed-off-by: Greg Kroah-Hartman

zsmalloc: collapse internal .h into .c

2012-08-14T02:34:24+00:00

The patch collapses in the internal zsmalloc_int.h into
the zsmalloc-main.c file.

This is done in preparation for the promotion to mm/ where
separate internal headers are discouraged.

Signed-off-by: Seth Jennings 
Signed-off-by: Minchan Kim 
Acked-by: Nitin Gupta 
Reviewed-by: Konrad Rzeszutek Wilk 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: add page table mapping method

2012-08-14T02:28:09+00:00

This patchset provides page mapping via the page table.
On some archs, most notably ARM, this method has been
demonstrated to be faster than copying.

The logic controlling the method selection (copy vs page table)
is controlled by the definition of USE_PGTABLE_MAPPING which
is/can be defined for any arch that performs better with page
table mapping.

Signed-off-by: Seth Jennings 
Acked-by: Minchan Kim 
Signed-off-by: Greg Kroah-Hartman

staging: zsmalloc: prevent mappping in interrupt context

2012-08-14T02:28:08+00:00

Because we use per-cpu mapping areas shared among the
pools/users, we can't allow mapping in interrupt context
because it can corrupt another users mappings.

Signed-off-by: Seth Jennings 
Acked-by: Minchan Kim 
Signed-off-by: Greg Kroah-Hartman