GitHub/LineageOS/G12/android_kernel_amlogic_linux-4.9.git
10 years agomm/rmap: recompute pgoff for huge page
Joonsoo Kim [Tue, 21 Jan 2014 23:49:43 +0000 (15:49 -0800)]
mm/rmap: recompute pgoff for huge page

Rmap traversing is used in five different cases, try_to_unmap(),
try_to_munlock(), page_referenced(), page_mkclean() and
remove_migration_ptes().  Each one implements its own traversing
functions for the cases, anon, file, ksm, respectively.  These cause
lots of duplications and cause maintenance overhead.  They also make
codes being hard to understand and error-prone.  One example is hugepage
handling.  There is a code to compute hugepage offset correctly in
try_to_unmap_file(), but, there isn't a code to compute hugepage offset
in rmap_walk_file().  These are used pairwise in migration context, but
we missed to modify pairwise.

To overcome these drawbacks, we should unify these through one unified
function.  I decide rmap_walk() as main function since it has no
unnecessity.  And to control behavior of rmap_walk(), I introduce struct
rmap_walk_control having some function pointers.  These makes
rmap_walk() working for their specific needs.

This patchset remove a lot of duplicated code as you can see in below
short-stat and kernel text size also decrease slightly.

   text    data     bss     dec     hex filename
  10640       1      16   10657    29a1 mm/rmap.o
  10047       1      16   10064    2750 mm/rmap.o

  13823     705    8288   22816    5920 mm/ksm.o
  13199     705    8288   22192    56b0 mm/ksm.o

This patch (of 9):

We have to recompute pgoff if the given page is huge, since result based
on HPAGE_SIZE is not approapriate for scanning the vma interval tree, as
shown by commit 36e4f20af833 ("hugetlb: do not use
vma_hugecache_offset() for vma_prio_tree_foreach") and commit 369a713e
("rmap: recompute pgoff for unmapping huge page").

To handle both the cases, normal page for page cache and hugetlb page,
by same way, we can use compound_page().  It returns 0 on non-compound
page and it also returns proper value on compound page.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Hugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemcg: make memcg_update_cache_sizes() static
Vladimir Davydov [Tue, 21 Jan 2014 23:49:42 +0000 (15:49 -0800)]
memcg: make memcg_update_cache_sizes() static

This function is not used outside of memcontrol.c so make it static.

Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemcg: fix kmem_account_flags check in memcg_can_account_kmem()
Vladimir Davydov [Tue, 21 Jan 2014 23:49:41 +0000 (15:49 -0800)]
memcg: fix kmem_account_flags check in memcg_can_account_kmem()

We should start kmem accounting for a memory cgroup only after both its
kmem limit is set (KMEM_ACCOUNTED_ACTIVE) and related call sites are
patched (KMEM_ACCOUNTED_ACTIVATED).  Currently memcg_can_account_kmem()
allows kmem accounting even if only one of the conditions is true.  Fix
it.

This means that a page might get charged by memcg_kmem_newpage_charge
which would see its static key patched already but
memcg_kmem_commit_charge would still see it unpatched and so the charge
won't be committed.  The result would be charge inconsistency
(page_cgroup not marked as PageCgroupUsed) and the charge would leak
because __memcg_kmem_uncharge_pages would ignore it.

[mhocko@suse.cz: augment changelog]
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Glauber Costa <glommer@parallels.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agox86, numa, acpi, memory-hotplug: make movable_node have higher priority
Tang Chen [Tue, 21 Jan 2014 23:49:38 +0000 (15:49 -0800)]
x86, numa, acpi, memory-hotplug: make movable_node have higher priority

If users specify the original movablecore=nn@ss boot option, the kernel
will arrange [ss, ss+nn) as ZONE_MOVABLE.  The kernelcore=nn@ss boot
option is similar except it specifies ZONE_NORMAL ranges.

Now, if users specify "movable_node" in kernel commandline, the kernel
will arrange hotpluggable memory in SRAT as ZONE_MOVABLE.  And if users
do this, all the other movablecore=nn@ss and kernelcore=nn@ss options
should be ignored.

For those who don't want this, just specify nothing.  The kernel will
act as before.

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemblock, mem_hotplug: make memblock skip hotpluggable regions if needed
Tang Chen [Tue, 21 Jan 2014 23:49:35 +0000 (15:49 -0800)]
memblock, mem_hotplug: make memblock skip hotpluggable regions if needed

Linux kernel cannot migrate pages used by the kernel.  As a result,
hotpluggable memory used by the kernel won't be able to be hot-removed.
To solve this problem, the basic idea is to prevent memblock from
allocating hotpluggable memory for the kernel at early time, and arrange
all hotpluggable memory in ACPI SRAT(System Resource Affinity Table) as
ZONE_MOVABLE when initializing zones.

In the previous patches, we have marked hotpluggable memory regions with
MEMBLOCK_HOTPLUG flag in memblock.memory.

In this patch, we make memblock skip these hotpluggable memory regions
in the default top-down allocation function if movable_node boot option
is specified.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoacpi, numa, mem_hotplug: mark all nodes the kernel resides un-hotpluggable
Tang Chen [Tue, 21 Jan 2014 23:49:32 +0000 (15:49 -0800)]
acpi, numa, mem_hotplug: mark all nodes the kernel resides un-hotpluggable

At very early time, the kernel have to use some memory such as loading
the kernel image.  We cannot prevent this anyway.  So any node the
kernel resides in should be un-hotpluggable.

Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoacpi, numa, mem_hotplug: mark hotpluggable memory in memblock
Tang Chen [Tue, 21 Jan 2014 23:49:29 +0000 (15:49 -0800)]
acpi, numa, mem_hotplug: mark hotpluggable memory in memblock

When parsing SRAT, we know that which memory area is hotpluggable.  So we
invoke function memblock_mark_hotplug() introduced by previous patch to
mark hotpluggable memory in memblock.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemblock: make memblock_set_node() support different memblock_type
Tang Chen [Tue, 21 Jan 2014 23:49:26 +0000 (15:49 -0800)]
memblock: make memblock_set_node() support different memblock_type

[sfr@canb.auug.org.au: fix powerpc build]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemblock, mem_hotplug: introduce MEMBLOCK_HOTPLUG flag to mark hotpluggable regions
Tang Chen [Tue, 21 Jan 2014 23:49:23 +0000 (15:49 -0800)]
memblock, mem_hotplug: introduce MEMBLOCK_HOTPLUG flag to mark hotpluggable regions

In find_hotpluggable_memory, once we find out a memory region which is
hotpluggable, we want to mark them in memblock.memory.  So that we could
control memblock allocator not to allocte hotpluggable memory for the
kernel later.

To achieve this goal, we introduce MEMBLOCK_HOTPLUG flag to indicate the
hotpluggable memory regions in memblock and a function
memblock_mark_hotplug() to mark hotpluggable memory if we find one.

[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Liu Jiang <jiang.liu@huawei.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomemblock, numa: introduce flags field into memblock
Tang Chen [Tue, 21 Jan 2014 23:49:20 +0000 (15:49 -0800)]
memblock, numa: introduce flags field into memblock

There is no flag in memblock to describe what type the memory is.
Sometimes, we may use memblock to reserve some memory for special usage.
And we want to know what kind of memory it is.  So we need a way to

In hotplug environment, we want to reserve hotpluggable memory so the
kernel won't be able to use it.  And when the system is up, we have to
free these hotpluggable memory to buddy.  So we need to mark these
memory first.

In order to do so, we need to mark out these special memory in memblock.
In this patch, we introduce a new "flags" member into memblock_region:

   struct memblock_region {
           phys_addr_t base;
           phys_addr_t size;
           unsigned long flags; /* This is new. */
   #ifdef CONFIG_HAVE_MEMBLOCK_NODE_MAP
           int nid;
   #endif
   };

This patch does the following things:
1) Add "flags" member to memblock_region.
2) Modify the following APIs' prototype:
memblock_add_region()
memblock_insert_region()
3) Add memblock_reserve_region() to support reserve memory with flags, and keep
   memblock_reserve()'s prototype unmodified.
4) Modify other APIs to support flags, but keep their prototype unmodified.

The idea is from Wen Congyang <wency@cn.fujitsu.com> and Liu Jiang <jiang.liu@huawei.com>.

Suggested-by: Wen Congyang <wency@cn.fujitsu.com>
Suggested-by: Liu Jiang <jiang.liu@huawei.com>
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Reviewed-by: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: "Rafael J . Wysocki" <rjw@sisk.pl>
Cc: Chen Tang <imtangchen@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiang Liu <jiang.liu@huawei.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Len Brown <lenb@kernel.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Prarit Bhargava <prarit@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Taku Izumi <izumi.taku@jp.fujitsu.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Thomas Renninger <trenn@suse.de>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/memblock: debug: correct displaying of upper memory boundary
Grygorii Strashko [Tue, 21 Jan 2014 23:49:17 +0000 (15:49 -0800)]
mm/memblock: debug: correct displaying of upper memory boundary

Current memblock APIs don't work on 32 PAE or LPAE extension arches
where the physical memory start address beyond 4GB.  The problem was
discussed here [3] where Tejun, Yinghai(thanks) proposed a way forward
with memblock interfaces.  Based on the proposal, this series adds
necessary memblock interfaces and convert the core kernel code to use
them.  Architectures already converted to NO_BOOTMEM use these new
interfaces and other which still uses bootmem, these new interfaces just
fallback to exiting bootmem APIs.

So no functional change in behavior.  In long run, once all the
architectures moves to NO_BOOTMEM, we can get rid of bootmem layer
completely.  This is one step to remove the core code dependency with
bootmem and also gives path for architectures to move away from bootmem.

Testing is done on ARM architecture with 32 bit ARM LAPE machines with
normal as well sparse(faked) memory model.

This patch (of 23):

When debugging is enabled (cmdline has "memblock=debug") the memblock
will display upper memory boundary per each allocated/freed memory range
wrongly.  For example:

 memblock_reserve: [0x0000009e7e8000-0x0000009e7ed000] _memblock_early_alloc_try_nid_nopanic+0xfc/0x12c

The 0x0000009e7ed000 is displayed instead of 0x0000009e7ecfff

Hence, correct this by changing formula used to calculate upper memory
boundary to (u64)base + size - 1 instead of (u64)base + size everywhere
in the debug messages.

Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Acked-by: Tejun Heo <tj@kernel.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/mlock: prepare params outside critical region
Davidlohr Bueso [Tue, 21 Jan 2014 23:49:16 +0000 (15:49 -0800)]
mm/mlock: prepare params outside critical region

All mlock related syscalls prepare lock limits, lengths and start
parameters with the mmap_sem held.  Move this logic outside of the
critical region.  For the case of mlock, continue incrementing the
amount already locked by mm->locked_vm with the rwsem taken.

Signed-off-by: Davidlohr Bueso <davidlohr@hp.com>
Cc: Rik van Riel <riel@redhat.com>
Reviewed-by: Michel Lespinasse <walken@google.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/mmap.c: add mlock_future_check() helper
Davidlohr Bueso [Tue, 21 Jan 2014 23:49:15 +0000 (15:49 -0800)]
mm/mmap.c: add mlock_future_check() helper

Both do_brk and do_mmap_pgoff verify that we are actually capable of
locking future pages if the corresponding VM_LOCKED flags are used.
Encapsulate this logic into a single mlock_future_check() helper
function.

Signed-off-by: Davidlohr Bueso <davidlohr@hp.com>
Cc: Rik van Riel <riel@redhat.com>
Reviewed-by: Michel Lespinasse <walken@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: add overcommit_kbytes sysctl variable
Jerome Marchand [Tue, 21 Jan 2014 23:49:14 +0000 (15:49 -0800)]
mm: add overcommit_kbytes sysctl variable

Some applications that run on HPC clusters are designed around the
availability of RAM and the overcommit ratio is fine tuned to get the
maximum usage of memory without swapping.  With growing memory, the
1%-of-all-RAM grain provided by overcommit_ratio has become too coarse
for these workload (on a 2TB machine it represents no less than 20GB).

This patch adds the new overcommit_kbytes sysctl variable that allow a
much finer grain.

[akpm@linux-foundation.org: coding-style fixes]
[akpm@linux-foundation.org: fix nommu build]
Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm, show_mem: remove SHOW_MEM_FILTER_PAGE_COUNT
Mel Gorman [Tue, 21 Jan 2014 23:49:13 +0000 (15:49 -0800)]
mm, show_mem: remove SHOW_MEM_FILTER_PAGE_COUNT

Commit 4b59e6c47309 ("mm, show_mem: suppress page counts in
non-blockable contexts") introduced SHOW_MEM_FILTER_PAGE_COUNT to
suppress PFN walks on large memory machines.  Commit c78e93630d15 ("mm:
do not walk all of system memory during show_mem") avoided a PFN walk in
the generic show_mem helper which removes the requirement for
SHOW_MEM_FILTER_PAGE_COUNT in that case.

This patch removes PFN walkers from the arch-specific implementations
that report on a per-node or per-zone granularity.  ARM and unicore32
still do a PFN walk as they report memory usage on each bank which is a
much finer granularity where the debugging information may still be of
use.  As the remaining arches doing PFN walks have relatively small
amounts of memory, this patch simply removes SHOW_MEM_FILTER_PAGE_COUNT.

[akpm@linux-foundation.org: fix parisc]
Signed-off-by: Mel Gorman <mgorman@suse.de>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: James Bottomley <jejb@parisc-linux.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/vmalloc: interchage the implementation of vmalloc_to_{pfn,page}
Jianyu Zhan [Tue, 21 Jan 2014 23:49:12 +0000 (15:49 -0800)]
mm/vmalloc: interchage the implementation of vmalloc_to_{pfn,page}

Currently we are implementing vmalloc_to_pfn() as a wrapper around
vmalloc_to_page(), which is implemented as follow:

 1. walks the page talbes to generates the corresponding pfn,
 2. then converts the pfn to struct page,
 3. returns it.

And vmalloc_to_pfn() re-wraps vmalloc_to_page() to get the pfn.

This seems too circuitous, so this patch reverses the way: implement
vmalloc_to_page() as a wrapper around vmalloc_to_pfn().  This makes
vmalloc_to_pfn() and vmalloc_to_page() slightly more efficient.

No functional change.

Signed-off-by: Jianyu Zhan <nasa4836@gmail.com>
Cc: Vladimir Murzin <murzin.v@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm, mempolicy: remove unneeded functions for UMA configs
David Rientjes [Tue, 21 Jan 2014 23:49:10 +0000 (15:49 -0800)]
mm, mempolicy: remove unneeded functions for UMA configs

Mempolicies only exist for CONFIG_NUMA configurations.  Therefore, a
certain class of functions are unneeded in configurations where
CONFIG_NUMA is disabled such as functions that duplicate existing
mempolicies, lookup existing policies, set certain mempolicy traits, or
test mempolicies for certain attributes.

Remove the unneeded functions so that any future callers get a compile-
time error and protect their code with CONFIG_NUMA as required.

Signed-off-by: David Rientjes <rientjes@google.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/hugetlb.c: call MMU notifiers when copying a hugetlb page range
Andreas Sandberg [Tue, 21 Jan 2014 23:49:09 +0000 (15:49 -0800)]
mm/hugetlb.c: call MMU notifiers when copying a hugetlb page range

When copy_hugetlb_page_range() is called to copy a range of hugetlb
mappings, the secondary MMUs are not notified if there is a protection
downgrade, which breaks COW semantics in KVM.

This patch adds the necessary MMU notifier calls.

Signed-off-by: Andreas Sandberg <andreas@sandberg.pp.se>
Acked-by: Steve Capper <steve.capper@linaro.org>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm, memory-failure: fix typo in me_pagecache_dirty()
Zhi Yong Wu [Tue, 21 Jan 2014 23:49:08 +0000 (15:49 -0800)]
mm, memory-failure: fix typo in me_pagecache_dirty()

[akpm@linux-foundation.org: s/cache/pagecache/]
Signed-off-by: Zhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: create a separate slab for page->ptl allocation
Kirill A. Shutemov [Tue, 21 Jan 2014 23:49:07 +0000 (15:49 -0800)]
mm: create a separate slab for page->ptl allocation

If DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC are enabled spinlock_t on x86_64
is 72 bytes.  For page->ptl they will be allocated from kmalloc-96 slab,
so we loose 24 on each.  An average system can easily allocate few tens
thousands of page->ptl and overhead is significant.

Let's create a separate slab for page->ptl allocation to solve this.

To make sure that it really works this time, some numbers from my test
machine (just booted, no load):

Before:
  # grep '^\(kmalloc-96\|page->ptl\)' /proc/slabinfo
  kmalloc-96         31987  32190    128   30    1 : tunables  120   60    8 : slabdata   1073   1073     92
After:
  # grep '^\(kmalloc-96\|page->ptl\)' /proc/slabinfo
  page->ptl          27516  28143     72   53    1 : tunables  120   60    8 : slabdata    531    531      9
  kmalloc-96          3853   5280    128   30    1 : tunables  120   60    8 : slabdata    176    176      0

Note that the patch is useful not only for debug case, but also for
PREEMPT_RT, where spinlock_t is always bloated.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: get rid of unnecessary pageblock scanning in setup_zone_migrate_reserve
Yasuaki Ishimatsu [Tue, 21 Jan 2014 23:49:06 +0000 (15:49 -0800)]
mm: get rid of unnecessary pageblock scanning in setup_zone_migrate_reserve

Yasuaki Ishimatsu reported memory hot-add spent more than 5 _hours_ on
9TB memory machine since onlining memory sections is too slow.  And we
found out setup_zone_migrate_reserve spent >90% of the time.

The problem is, setup_zone_migrate_reserve scans all pageblocks
unconditionally, but it is only necessary if the number of reserved
block was reduced (i.e.  memory hot remove).

Moreover, maximum MIGRATE_RESERVE per zone is currently 2.  It means
that the number of reserved pageblocks is almost always unchanged.

This patch adds zone->nr_migrate_reserve_block to maintain the number of
MIGRATE_RESERVE pageblocks and it reduces the overhead of
setup_zone_migrate_reserve dramatically.  The following table shows time
of onlining a memory section.

  Amount of memory     | 128GB | 192GB | 256GB|
  ---------------------------------------------
  linux-3.12           |  23.9 |  31.4 | 44.5 |
  This patch           |   8.3 |   8.3 |  8.6 |
  Mel's proposal patch |  10.9 |  19.2 | 31.3 |
  ---------------------------------------------
                                   (millisecond)

  128GB : 4 nodes and each node has 32GB of memory
  192GB : 6 nodes and each node has 32GB of memory
  256GB : 8 nodes and each node has 32GB of memory

  (*1) Mel proposed his idea by the following threads.
       https://lkml.org/lkml/2013/10/30/272

[akpm@linux-foundation.org: tweak comment]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Reported-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Tested-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Cc: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years ago/proc/meminfo: provide estimated available memory
Rik van Riel [Tue, 21 Jan 2014 23:49:05 +0000 (15:49 -0800)]
/proc/meminfo: provide estimated available memory

Many load balancing and workload placing programs check /proc/meminfo to
estimate how much free memory is available.  They generally do this by
adding up "free" and "cached", which was fine ten years ago, but is
pretty much guaranteed to be wrong today.

It is wrong because Cached includes memory that is not freeable as page
cache, for example shared memory segments, tmpfs, and ramfs, and it does
not include reclaimable slab memory, which can take up a large fraction
of system memory on mostly idle systems with lots of files.

Currently, the amount of memory that is available for a new workload,
without pushing the system into swap, can be estimated from MemFree,
Active(file), Inactive(file), and SReclaimable, as well as the "low"
watermarks from /proc/zoneinfo.

However, this may change in the future, and user space really should not
be expected to know kernel internals to come up with an estimate for the
amount of free memory.

It is more convenient to provide such an estimate in /proc/meminfo.  If
things change in the future, we only have to change it in one place.

Signed-off-by: Rik van Riel <riel@redhat.com>
Reported-by: Erik Mouw <erik.mouw_2@nxp.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: thp: turn compound_head() into BUG_ON(!PageTail) in get_huge_page_tail()
Oleg Nesterov [Tue, 21 Jan 2014 23:49:04 +0000 (15:49 -0800)]
mm: thp: turn compound_head() into BUG_ON(!PageTail) in get_huge_page_tail()

get_huge_page_tail()->compound_head() looks confusing.  Every caller
must check PageTail(page), otherwise atomic_inc(&page->_mapcount) is
simply wrong if this page is compound-trans-head.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Dave Jones <davej@redhat.com>
Cc: Darren Hart <dvhart@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@suse.de>
Acked-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: thp: __get_page_tail_foll() can use get_huge_page_tail()
Oleg Nesterov [Tue, 21 Jan 2014 23:49:02 +0000 (15:49 -0800)]
mm: thp: __get_page_tail_foll() can use get_huge_page_tail()

Cleanup. Change __get_page_tail_foll() to use get_huge_page_tail()
to avoid the code duplication.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Dave Jones <davej@redhat.com>
Cc: Darren Hart <dvhart@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@suse.de>
Acked-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/hugetlb.c: defer PageHeadHuge() symbol export
Andrea Arcangeli [Tue, 21 Jan 2014 23:49:01 +0000 (15:49 -0800)]
mm/hugetlb.c: defer PageHeadHuge() symbol export

No actual need of it. So keep it internal.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/swap.c: reorganize put_compound_page()
Andrew Morton [Tue, 21 Jan 2014 23:48:59 +0000 (15:48 -0800)]
mm/swap.c: reorganize put_compound_page()

Tweak it so save a tab stop, make code layout slightly less nutty.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm/hugetlb.c: simplify PageHeadHuge() and PageHuge()
Andrew Morton [Tue, 21 Jan 2014 23:48:57 +0000 (15:48 -0800)]
mm/hugetlb.c: simplify PageHeadHuge() and PageHuge()

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: hugetlbfs: use __compound_tail_refcounted in __get_page_tail too
Andrea Arcangeli [Tue, 21 Jan 2014 23:48:56 +0000 (15:48 -0800)]
mm: hugetlbfs: use __compound_tail_refcounted in __get_page_tail too

Also remove hugetlb.h which isn't needed anymore as PageHeadHuge is
handled in mm.h.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: tail page refcounting optimization for slab and hugetlbfs
Andrea Arcangeli [Tue, 21 Jan 2014 23:48:54 +0000 (15:48 -0800)]
mm: tail page refcounting optimization for slab and hugetlbfs

This skips the _mapcount mangling for slab and hugetlbfs pages.

The main trouble in doing this is to guarantee that PageSlab and
PageHeadHuge remains constant for all get_page/put_page run on the tail
of slab or hugetlbfs compound pages.  Otherwise if they're set during
get_page but not set during put_page, the _mapcount of the tail page
would underflow.

PageHeadHuge will remain true until the compound page is released and
enters the buddy allocator so it won't risk to change even if the tail
page is the last reference left on the page.

PG_slab instead is cleared before the slab frees the head page with
put_page, so if the tail pin is released after the slab freed the page,
we would have a problem.  But in the slab case the tail pin cannot be
the last reference left on the page.  This is because the slab code is
free to reuse the compound page after a kfree/kmem_cache_free without
having to check if there's any tail pin left.  In turn all tail pins
must be always released while the head is still pinned by the slab code
and so we know PG_slab will be still set too.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Reviewed-by: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: thp: optimize compound_trans_huge
Andrea Arcangeli [Tue, 21 Jan 2014 23:48:52 +0000 (15:48 -0800)]
mm: thp: optimize compound_trans_huge

Currently we don't clobber page_tail->first_page during split_huge_page,
so compound_trans_head can be set to compound_head without adverse
effects, and this mostly optimizes away a smp_rmb.

It looks worthwhile to keep around the implementation that doesn't relay
on page_tail->first_page not to be clobbered, because it would be
necessary if we'll decide to enforce page->private to zero at all times
whenever PG_private is not set, also for anonymous pages.  For anonymous
pages enforcing such an invariant doesn't matter as anonymous pages
don't use page->private so we can get away with this microoptimization.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: hugetlbfs: move the put/get_page slab and hugetlbfs optimization in a faster...
Andrea Arcangeli [Tue, 21 Jan 2014 23:48:51 +0000 (15:48 -0800)]
mm: hugetlbfs: move the put/get_page slab and hugetlbfs optimization in a faster path

We don't actually need a reference on the head page in the slab and
hugetlbfs paths, as long as we add a smp_rmb() which should be faster
than get_page_unless_zero.

[akpm@linux-foundation.org: fix typo in comment]
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: hugetlb: use get_page_foll() in follow_hugetlb_page()
Andrea Arcangeli [Tue, 21 Jan 2014 23:48:49 +0000 (15:48 -0800)]
mm: hugetlb: use get_page_foll() in follow_hugetlb_page()

get_page_foll() is more optimal and is always safe to use under the PT
lock.  More so for hugetlbfs as there's no risk of race conditions with
split_huge_page regardless of the PT lock.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Tested-by: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Pravin Shelar <pshelar@nicira.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: hugetlbfs: Add some VM_BUG_ON()s to catch non-hugetlbfs pages
Dave Hansen [Tue, 21 Jan 2014 23:48:48 +0000 (15:48 -0800)]
mm: hugetlbfs: Add some VM_BUG_ON()s to catch non-hugetlbfs pages

Dave Jiang reported that he was seeing oopses when running NUMA systems
and default_hugepagesz=1G.  I traced the issue down to
migrate_page_copy() trying to use the same code for hugetlb pages and
transparent hugepages.  It should not have been trying to pass thp pages
in there.

So, add some VM_BUG_ON()s for the next hapless VM developer that tries
the same thing.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Tested-by: Dave Jiang <dave.jiang@intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: Make {,set}page_address() static inline if WANT_PAGE_VIRTUAL
Geert Uytterhoeven [Tue, 21 Jan 2014 23:48:47 +0000 (15:48 -0800)]
mm: Make {,set}page_address() static inline if WANT_PAGE_VIRTUAL

{,set}page_address() are macros if WANT_PAGE_VIRTUAL.  If
!WANT_PAGE_VIRTUAL, they're plain C functions.

If someone calls them with a void *, this pointer is auto-converted to
struct page * if !WANT_PAGE_VIRTUAL, but causes a build failure on
architectures using WANT_PAGE_VIRTUAL (arc, m68k and sparc64):

  drivers/md/bcache/bset.c: In function `__btree_sort':
  drivers/md/bcache/bset.c:1190: warning: dereferencing `void *' pointer
  drivers/md/bcache/bset.c:1190: error: request for member `virtual' in something not a structure or union

Convert them to static inline functions to fix this.  There are already
plenty of users of struct page members inside <linux/mm.h>, so there's
no reason to keep them as macros.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Tested-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofs/ramfs: don't use module_init for non-modular core code
Paul Gortmaker [Tue, 21 Jan 2014 23:48:46 +0000 (15:48 -0800)]
fs/ramfs: don't use module_init for non-modular core code

The ramfs is always built in.  It will never be modular, so using
module_init as an alias for __initcall is rather misleading.

Fix this up now, so that we can relocate module_init from init.h into
module.h in the future.  If we don't do this, we'd have to add module.h
to obviously non-modular code, and that would be a worse thing.

Note that direct use of __initcall is discouraged, vs.  one of the
priority categorized subgroups.  As __initcall gets mapped onto
device_initcall, our use of fs_initcall (which makes sense for fs code)
will thus change this registration from level 6-device to level 5-fs
(i.e. slightly earlier).  However no observable impact of that small
difference has been observed during testing, or is expected.

Also note that this change uncovers a missing semicolon bug in the
registration of the initcall.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofs/super.c: fix WARN on alloc_super() fail path
Vladimir Davydov [Tue, 21 Jan 2014 23:48:45 +0000 (15:48 -0800)]
fs/super.c: fix WARN on alloc_super() fail path

On fail path alloc_super() calls destroy_super(), which issues a warning
if the sb's s_mounts list is not empty, in particular if it has not been
initialized.  That said s_mounts must be initialized in alloc_super()
before any possible failure, but currently it is initialized close to
the end of the function leading to a useless warning dumped to log if
either percpu_counter_init() or list_lru_init() fails.  Let's fix this.

Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofs/read_write.c:compat_readv(): remove bogus area verify
Corey Minyard [Tue, 21 Jan 2014 23:48:44 +0000 (15:48 -0800)]
fs/read_write.c:compat_readv(): remove bogus area verify

The compat_do_readv_writev() function was doing a verify_area on the
incoming iov, but the nr_segs value is not checked.  If someone passes
in a -1 for nr_segs, for instance, the function should return an EINVAL.
However, it returns a EFAULT because the verify_area fails because it is
checking an array of size MAX_UINT.  The check is bogus, anyway, because
the next check, compat_rw_copy_check_uvector(), will do all the
necessary checking, anyway.  The non-compat do_readv_writev() function
doesn't do this check, so I think it's safe to just remove the code.

Signed-off-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofs/compat_ioctl.c: fix an underflow issue (harmless)
Dan Carpenter [Tue, 21 Jan 2014 23:48:43 +0000 (15:48 -0800)]
fs/compat_ioctl.c: fix an underflow issue (harmless)

We cap "nmsgs" at I2C_RDRW_IOCTL_MAX_MSGS (42) but the current code
allows negative values.  It's harmless but it makes my static checker
upset so I've made nsmgs unsigned.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoposix_acl: uninlining
Andrew Morton [Tue, 21 Jan 2014 23:48:42 +0000 (15:48 -0800)]
posix_acl: uninlining

Uninline vast tracts of nested inline functions in
include/linux/posix_acl.h.

This reduces the text+data+bss size of x86_64 allyesconfig vmlinux by
8026 bytes.

The patch also regularises the positioning of the EXPORT_SYMBOLs in
posix_acl.c.

Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: J. Bruce Fields <bfields@fieldses.org>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Tested-by: Benny Halevy <bhalevy@primarydata.com>
Cc: Benny Halevy <bhalevy@panasas.com>
Cc: Andreas Gruenbacher <agruen@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoarch/sh/kernel/kgdb.c: add missing #include <linux/sched.h>
Wanlong Gao [Tue, 21 Jan 2014 23:48:41 +0000 (15:48 -0800)]
arch/sh/kernel/kgdb.c: add missing #include <linux/sched.h>

  arch/sh/kernel/kgdb.c: In function 'sleeping_thread_to_gdb_regs':
  arch/sh/kernel/kgdb.c:225:32: error: implicit declaration of function 'task_stack_page' [-Werror=implicit-function-declaration]
  arch/sh/kernel/kgdb.c:242:23: error: dereferencing pointer to incomplete type
  arch/sh/kernel/kgdb.c:243:22: error: dereferencing pointer to incomplete type
  arch/sh/kernel/kgdb.c: In function 'singlestep_trap_handler':
  arch/sh/kernel/kgdb.c:310:27: error: 'SIGTRAP' undeclared (first use in this function)
  arch/sh/kernel/kgdb.c:310:27: note: each undeclared identifier is reported only once for each function it appears in

This was introduced by commit 16559ae48c76 ("kgdb: remove #include
<linux/serial_8250.h> from kgdb.h").

[geert@linux-m68k.org: reworded and reformatted]
Signed-off-by: Wanlong Gao <gaowanlong@cn.fujitsu.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: fix NULL pointer dereference when dismount and ocfs2rec simultaneously
Yiwen Jiang [Tue, 21 Jan 2014 23:48:39 +0000 (15:48 -0800)]
ocfs2: fix NULL pointer dereference when dismount and ocfs2rec simultaneously

2 nodes cluster, say Node A and Node B, mount the same ocfs2 volume, and
create a file 1.

Node A Node B
open 1, get open lock
                        rm 1, and then add 1 to orphan_dir
storage link down,
o2hb_write_timeout
->o2quo_disk_timeout
->emergency_restart
                        at the moment, Node B dismount and do
ocfs2rec simultaneously
                        1) ocfs2_dismount_volume
->ocfs2_recovery_exit
->wait_event(osb->recovery_event)
->flush_workqueue(ocfs2_wq)
2) ocfs2rec
->queue_work(&journal->j_recovery_work)
                        ->ocfs2_recover_orphans
->ocfs2_commit_truncate
                        ->queue_delayed_work(&osb->osb_truncate_log_wq)

In ocfs2_recovery_exit, it flushes workqueue and then releases system
inodes.  When doing ocfs2rec, it will call ocfs2_flush_truncate_log
which will try to get sys_root_inode, and NULL pointer dereference
occurs.

Signed-off-by: Yiwen Jiang <jiangyiwen@huawei.com>
Signed-off-by: joyce <xuejiufei@huawei.com>
Signed-off-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: punch hole should return EINVAL if the length argument in ioctl is negative
Tariq Saeed [Tue, 21 Jan 2014 23:48:38 +0000 (15:48 -0800)]
ocfs2: punch hole should return EINVAL if the length argument in ioctl is negative

An unreserve space ioctl OCFS2_IOC_UNRESVSP/64 should reject a negative
length.

Orabug:14789508

Signed-off-by: Tariq Saseed <tariq.x.saeed@oracle.com>
Signed-off-by: Srinivas Eeda <srinivas.eeda@oracle.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: fix sparse non static symbol warning
Wei Yongjun [Tue, 21 Jan 2014 23:48:37 +0000 (15:48 -0800)]
ocfs2: fix sparse non static symbol warning

Fixes the following sparse warning:

  fs/ocfs2/stack_user.c:930:32: warning:
   symbol 'ocfs2_ls_ops' was not declared. Should it be static?

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: adjust minlen with discard_granularity in the FITRIM ioctl
Jie Liu [Tue, 21 Jan 2014 23:48:36 +0000 (15:48 -0800)]
ocfs2: adjust minlen with discard_granularity in the FITRIM ioctl

Adjust minlen with discard_granularity for FITRIM ioctl(2) if the given
minimum size in bytes is less than it because, discard granularity is
used to tell us that the minimum size of extent that can be discarded by
the storage device.

This is inspired by ext4 commit 5c2ed62fd447 ("ext4: Adjust minlen with
discard_granularity in the FITRIM ioctl") from Lukas Czerner.

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: return EINVAL if the given range to discard is less than block size
Jie Liu [Tue, 21 Jan 2014 23:48:35 +0000 (15:48 -0800)]
ocfs2: return EINVAL if the given range to discard is less than block size

For FITRIM ioctl(2), we should not keep silence if the given range
length ls less than a block size as there is no data blocks would be
discareded.  Hence it should return EINVAL instead.  This issue can be
verified via xfstests/generic/288 which is used for FITRIM argument
handling tests.

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: return EOPNOTSUPP if the device does not support discard
Jie Liu [Tue, 21 Jan 2014 23:48:34 +0000 (15:48 -0800)]
ocfs2: return EOPNOTSUPP if the device does not support discard

For FITRIM ioctl(2), we should return EOPNOTSUPP to inform the user that
the storage device does not support discard if it is, otherwise return
success would confuse the user even though there is no free blocks were
trimmed at all.

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: remove redundant ocfs2_alloc_dinode_update_counts() and ocfs2_block_group_set_...
Younger Liu [Tue, 21 Jan 2014 23:48:33 +0000 (15:48 -0800)]
ocfs2: remove redundant ocfs2_alloc_dinode_update_counts() and ocfs2_block_group_set_bits()

ocfs2_alloc_dinode_update_counts() and ocfs2_block_group_set_bits() are
already provided in suballoc.c.  So, the same functions in
move_extents.c are not needed any more.

Declare the functions in suballoc.h and remove redundant functions in
move_extents.c.

Signed-off-by: Younger Liu <liuyiyang@hisense.com>
Cc: Younger Liu <younger.liucn@gmail.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: use the new DLM operation callbacks while requesting new lockspace
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:32 +0000 (15:48 -0800)]
ocfs2: use the new DLM operation callbacks while requesting new lockspace

Attempt to use the new DLM operations.  If it is not supported, use the
traditional ocfs2_controld.

To exchange ocfs2 versioning, we use the LVB of the version dlm lock.
It first attempts to take the lock in EX mode (non-blocking).  If
successful (which means it is the first mount), it writes the version
number and downconverts to PR lock.  If it is unsuccessful, it reads the
version from the lock.

If this becomes the standard (with o2cb as well), it could simplify
userspace tools to check if the filesystem is mounted on other nodes.

Dan: Since ocfs2_protocol_version are two u8 values, the additional
checks with LONG* don't make sense.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: framework for version LVB
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:25 +0000 (15:48 -0800)]
ocfs2: framework for version LVB

Use the native DLM locks for version control negotiation.  Most of the
framework is taken from gfs2/lock_dlm.c

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: pass ocfs2_cluster_connection to ocfs2_this_node
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:24 +0000 (15:48 -0800)]
ocfs2: pass ocfs2_cluster_connection to ocfs2_this_node

This is done to differentiate between using and not using controld and
use the connection information accordingly.

We need to be backward compatible.  So, we use a new enum
ocfs2_connection_type to identify when controld is used and when it is
not.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: shift allocation ocfs2_live_connection to user_connect()
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:23 +0000 (15:48 -0800)]
ocfs2: shift allocation ocfs2_live_connection to user_connect()

We perform this because the DLM recovery callbacks will require the
ocfs2_live_connection structure to record the node information when
dlm_new_lockspace() is updated (in the last patch of the series).

Before calling dlm_new_lockspace(), we need the structure ready for the
.recover_done() callback, which would set oc_this_node.  This is the
reason we allocate ocfs2_live_connection beforehand in user_connect().

[AKPM] rc initialization is not required because it assigned in case of
errors.  It will be cleared by compiler anyways.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reveiwed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: add DLM recovery callbacks
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:22 +0000 (15:48 -0800)]
ocfs2: add DLM recovery callbacks

These are the callbacks called by the fs/dlm code in case the membership
changes.  If there is a failure while/during calling any of these, the
DLM creates a new membership and relays to the rest of the nodes.

 - recover_prep() is called when DLM understands a node is down.
 - recover_slot() is called once all nodes have acknowledged
   recover_prep and recovery can begin.
 - recover_done() is called once the recovery is complete.  It returns
   the new membership.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: add clustername to cluster connection
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:21 +0000 (15:48 -0800)]
ocfs2: add clustername to cluster connection

This is an effort of removing ocfs2_controld.pcmk and getting ocfs2 DLM
handling up to the times with respect to DLM (>=4.0.1) and corosync
(2.3.x).  AFAIK, cman also is being phased out for a unified corosync
cluster stack.

fs/dlm performs all the functions with respect to fencing and node
management and provides the API's to do so for ocfs2.  For all future
references, DLM stands for fs/dlm code.

The advantages are:
 + No need to run an additional userspace daemon (ocfs2_controld)
 + No controld device handling and controld protocol
 + Shifting responsibilities of node management to DLM layer

For backward compatibility, we are keeping the controld handling code.
Once enough time has passed we can remove a significant portion of the
code.  This was tested by using the kernel with changes on older
unmodified tools.  The kernel used ocfs2_controld as expected, and
displayed the appropriate warning message.

This feature requires modification in the userspace ocfs2-tools.  The
changes can be found at: https://github.com/goldwynr/ocfs2-tools branch:
nocontrold Currently, not many checks are present in the userspace code,
but that would change soon.

This patch (of 6):

Add clustername to cluster connection.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoocfs2: remove versioning information
Goldwyn Rodrigues [Tue, 21 Jan 2014 23:48:20 +0000 (15:48 -0800)]
ocfs2: remove versioning information

The versioning information is confusing for end-users.  The numbers are
stuck at 1.5.0 when the tools version have moved to 1.8.2.  Remove the
versioning system in the OCFS2 modules and let the kernel version be the
guide to debug issues.

Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Sunil Mushran <sunil.mushran@gmail.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Acked-by: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoscore: remove "select HAVE_GENERIC_HARDIRQS" again
Geert Uytterhoeven [Tue, 21 Jan 2014 23:48:19 +0000 (15:48 -0800)]
score: remove "select HAVE_GENERIC_HARDIRQS" again

Commit 5fbbf8a1a934 ("Score: The commit is for compiling successfully.")
re-introduced "select HAVE_GENERIC_HARDIRQS" in v3.12-rc4, which had
just been removed in v3.12-rc1 by 0244ad004a5 ("Remove GENERIC_HARDIRQ
config option").

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Chen Liqin <liqin.linux@gmail.com>
Cc: Lennox Wu <lennox.wu@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agointel-iommu: fix off-by-one in pagetable freeing
Alex Williamson [Tue, 21 Jan 2014 23:48:18 +0000 (15:48 -0800)]
intel-iommu: fix off-by-one in pagetable freeing

dma_pte_free_level() has an off-by-one error when checking whether a pte
is completely covered by a range.  Take for example the case of
attempting to free pfn 0x0 - 0x1ff, ie.  512 entries covering the first
2M superpage.

The level_size() is 0x200 and we test:

  static void dma_pte_free_level(...
...

if (!(0 > 0 || 0x1ff < 0 + 0x200)) {
...
}

Clearly the 2nd test is true, which means we fail to take the branch to
clear and free the pagetable entry.  As a result, we're leaking
pagetables and failing to install new pages over the range.

This was found with a PCI device assigned to a QEMU guest using vfio-pci
without a VGA device present.  The first 1M of guest address space is
mapped with various combinations of 4K pages, but eventually the range
is entirely freed and replaced with a 2M contiguous mapping.
intel-iommu errors out with something like:

  ERROR: DMA PTE for vPFN 0x0 already set (to 5c2b8003 not 849c00083)

In this case 5c2b8003 is the pointer to the previous leaf page that was
neither freed nor cleared and 849c00083 is the superpage entry that
we're trying to replace it with.

Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofsnotify: remove pointless NULL initializers
Jan Kara [Tue, 21 Jan 2014 23:48:16 +0000 (15:48 -0800)]
fsnotify: remove pointless NULL initializers

We usually rely on the fact that struct members not specified in the
initializer are set to NULL.  So do that with fsnotify function pointers
as well.

Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Eric Paris <eparis@parisplace.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofsnotify: remove .should_send_event callback
Jan Kara [Tue, 21 Jan 2014 23:48:15 +0000 (15:48 -0800)]
fsnotify: remove .should_send_event callback

After removing event structure creation from the generic layer there is
no reason for separate .should_send_event and .handle_event callbacks.
So just remove the first one.

Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Eric Paris <eparis@parisplace.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agofsnotify: do not share events between notification groups
Jan Kara [Tue, 21 Jan 2014 23:48:14 +0000 (15:48 -0800)]
fsnotify: do not share events between notification groups

Currently fsnotify framework creates one event structure for each
notification event and links this event into all interested notification
groups.  This is done so that we save memory when several notification
groups are interested in the event.  However the need for event
structure shared between inotify & fanotify bloats the event structure
so the result is often higher memory consumption.

Another problem is that fsnotify framework keeps path references with
outstanding events so that fanotify can return open file descriptors
with its events.  This has the undesirable effect that filesystem cannot
be unmounted while there are outstanding events - a regression for
inotify compared to a situation before it was converted to fsnotify
framework.  For fanotify this problem is hard to avoid and users of
fanotify should kind of expect this behavior when they ask for file
descriptors from notified files.

This patch changes fsnotify and its users to create separate event
structure for each group.  This allows for much simpler code (~400 lines
removed by this patch) and also smaller event structures.  For example
on 64-bit system original struct fsnotify_event consumes 120 bytes, plus
additional space for file name, additional 24 bytes for second and each
subsequent group linking the event, and additional 32 bytes for each
inotify group for private data.  After the conversion inotify event
consumes 48 bytes plus space for file name which is considerably less
memory unless file names are long and there are several groups
interested in the events (both of which are uncommon).  Fanotify event
fits in 56 bytes after the conversion (fanotify doesn't care about file
names so its events don't have to have it allocated).  A win unless
there are four or more fanotify groups interested in the event.

The conversion also solves the problem with unmount when only inotify is
used as we don't have to grab path references for inotify events.

[hughd@google.com: fanotify: fix corruption preventing startup]
Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Eric Paris <eparis@parisplace.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoinotify: provide function for name length rounding
Jan Kara [Tue, 21 Jan 2014 23:48:13 +0000 (15:48 -0800)]
inotify: provide function for name length rounding

Rounding of name length when passing it to userspace was done in several
places.  Provide a function to do it and use it in all places.

Signed-off-by: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Eric Paris <eparis@parisplace.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agodma-debug: introduce debug_dma_assert_idle()
Dan Williams [Tue, 21 Jan 2014 23:48:12 +0000 (15:48 -0800)]
dma-debug: introduce debug_dma_assert_idle()

Record actively mapped pages and provide an api for asserting a given
page is dma inactive before execution proceeds.  Placing
debug_dma_assert_idle() in cow_user_page() flagged the violation of the
dma-api in the NET_DMA implementation (see commit 77873803363c "net_dma:
mark broken").

The implementation includes the capability to count, in a limited way,
repeat mappings of the same page that occur without an intervening
unmap.  This 'overlap' counter is limited to the few bits of tag space
in a radix tree.  This mechanism is added to mitigate false negative
cases where, for example, a page is dma mapped twice and
debug_dma_assert_idle() is called after the page is un-mapped once.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Vinod Koul <vinod.koul@intel.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMerge tag 'for-v3.14' of git://git.infradead.org/battery-2.6
Linus Torvalds [Tue, 21 Jan 2014 19:36:20 +0000 (11:36 -0800)]
Merge tag 'for-v3.14' of git://git.infradead.org/battery-2.6

Pull battery updates from Dmitry Eremin-Solenikov:
 "I'm picking up power supply maintainership from Anton Vorontov.  Could
  you please pull battery-2.6 git tree changes prepared for the v3.14
  release.

  Highlights:

   - Power supply notifier

   - Several drivers gained DT support

   - Added Maxim 14577 driver

   - Change of maintainer"

* tag 'for-v3.14' of git://git.infradead.org/battery-2.6:
  MAINTAINERS: Pick up power supply maintainership
  max17042_battery: Add IRQF_ONESHOT flag to use default irq handler
  gpio-charger: Support wakeup events
  power_supply: Add charger support for Maxim 14577
  dt: Binding documentation for isp1704 charger
  isp1704_charger: Add DT support
  charger-manager: of_cm_parse_desc() should be static
  bq2415x_charger: Add DT support
  power_supply: Add power_supply_get_by_phandle
  bq2415x_charger: Use power_supply notifier for automode
  power: reset: Add as3722 power-off driver
  mfd: AS3722: Add dt node properties for system power controller
  charger-manager: Support deivce tree in charger manager driver
  charger-manager: Modify the way of checking battery's temperature
  power_supply: Add power_supply notifier

10 years agoMerge tag 'mfd-3.14-1' of git://git.linaro.org/people/ljones/mfd
Linus Torvalds [Tue, 21 Jan 2014 18:58:17 +0000 (10:58 -0800)]
Merge tag 'mfd-3.14-1' of git://git.linaro.org/people/ljones/mfd

Pull MFD changes from Lee Jones:
 "New drivers
   - Samsung Maxim 14577; Micro USB, Regulator, IRQ Controller and
     Battery Charger
   - TI/National Semiconductor LP3943 I2C GPIO Expander and PWM
     Generator

  Existing driver adaptions
   - Expansion of Wolfson Arizona DSP and High-Pass filter controls
   - TI TWL6040 default Regmap support and Regcache addition/bypass
   - Some nice Smatch catch fixes
   - Conversion of TI OMAP-USB and TI TWL6030 to endian neutralness
   - ChromeOS EC timing (delay) adaptions and added dependency on OF
   - Many constifications of 'struct {mfd_cell,regmap_irq,et.al}'
   - Watchdog support added for NVIDIA AS3722
   - Convert functions to static in TI AM335x
   - Realigned previously defeated functionality in TI AM335x
   - IIO ADC-TSC concurrency dead-lock/timeout resolution
   - Addition of Power Management and Clock support for Samsung core
   - DEFINE_PCI_DEVICE_TABLE macro removal from MFD Subsystem
   - Greater use of irqdomain functionality in ST-E AB8500
   - Removal of 'include/linux/mfd/abx500/ab8500-gpio.h'
   - Wolfson WM831x PMIC Power Management changes s/poweroff/shutdown/
   - Device Tree documentation added for TI/Nat Semi LP3943
   - Version detection and voltage tables for TI TPS6586x PMIC devices
   - Simplification of Freescale MC13XXX (de-)initialisation routines
   - Clean-up and simplification of the Realtek parent driver
   - Added support for RTL8402 Realtek PCI-Express card reader
   - Resource leak fix for Maxim 77686
   - Possible suspend BUG() fix in OMAP USB TLL
   - Support for new Wolfson WM5110 Revision (D)
   - Testing of automatic assignment of of_node in mfd_add_device()
   - Reversion of the above when it started to cause issues
   - Remove legacy Platform Data from;
              TI TWL Core, Qualcomm SSBI and ST-E ABx500 Pinctrl
   - Clean-ups; tabbing issues, function name changes, 'drvdata = NULL'
              removal, unused uninitialised warning mitigation, error
              message clarity, removal of redundant/duplicate checks,
              licensing (GPL -> GPL2), coding consistency, duplicate
              function declaration, ret checks, commit corrections,
              redundant of_match_ptr() helper removal, spelling,
              #if-deffery removal and header guards name changes"

* tag 'mfd-3.14-1' of git://git.linaro.org/people/ljones/mfd: (78 commits)
  mfd: wm5110: Add register patch for rev D chip
  mfd: omap-usb-tll: Don't hold lock during pm_runtime_get/put_sync()
  gpio: lp3943: Remove redundant of_match_ptr helper
  mfd: sta2x11-mfd: Use named constants for pci_power_t values
  Documentation: mfd: Fix LDO index in s2mps11.txt
  mfd: Cleanup mfd-mcp-sa11x0.h header
  mfd: max8997: Use "IS_ENABLED(CONFIG_OF)" for DT code.
  mfd: twl6030: Fix endianness problem in IRQ handler
  mfd: sec-core: Add cells for S5M8767-clocks
  mfd: max14577: Remove redundant of_match_ptr helper
  mfd: twl6040: Fix sparse non static symbol warning
  mfd: Revert "mfd: Always assign of_node in mfd_add_device()"
  mfd: rtsx: Fix sparse non static symbol warning
  mfd: max77693: Set proper maximum register for MUIC regmap
  mfd: max77686: Fix regmap resource leak on driver remove
  mfd: Represent correct filenames in file headers
  mfd: rtsx: Add support for card reader rtl8402
  mfd: rtsx: Add set pull control macro and simplify rtl8411
  mfd: max8997: Enforce mfd_add_devices() return value check
  mfd: mc13xxx: Simplify probe() & remove()
  ...

10 years agoMerge tag 'sound-3.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Tue, 21 Jan 2014 18:26:23 +0000 (10:26 -0800)]
Merge tag 'sound-3.14-rc1' of git://git./linux/kernel/git/tiwai/sound

Pull sound updates from Takashi Iwai:
 "It was holiday season, so no wonder that there are little changes in
  framework level, although diffstat shows quite many changes spreaded
  over sound/* directories.  Most of changes are cleanups, code
  refactoring and fixes.

  Some highlights:
   - Removal of OSS sleep_on usages by Arnd
   - Simplified memalloc helper codes, drop obsoleted features; now it's
     built into PCM driver instead of an individual module
   - Warn if PCM buffer preallocation fails, which will show page
     allocation issues more clearly
   - Compress offload API updates for sample rates by Vinod
   - PCM glitch workaround on ctxfi emu20k1 by Sarah
   - Drop cs46xx DSP blobs, using firmware loader now
   - USB-audio quitks for Plantronics Gamecom 780, Creative VF0420, and
     Focusrite Saffire 6

  HD-audio specifics:
   - Standardize Kconfigs of HD-audio codec drivers; now "make
     localmodconfig" recognizes configs properly (finally!)
   - Parallel PM implementation by Mengdong
   - BayleyBay/ValleyView2 board fixups
   - Broadwell audio support
   - Runtime PM improvement (PantherPoint, etc)
   - Quirks: Dell subwooer, Gigabyte mobo jack detection oddity, Dell
     AiO click noise fixes, Dell headset mic fixes, etc
   - Automatic bind with HDMI codec parser without generic parser
   - More AD codec fixes (since 3.12 regression) including the automatic
     stereo mix support
   - Common Thinkpad ACPI helper for Realtek and Conexant codecs

  ASoC specifics:
   - Update to the generic DMA code to support deferred probe and
     managed resources
   - New drivers for BCM2835 (used in Raspberry Pi), Tegra with MAX98090
     and Analog Devices AXI I2S and S/PDIF controller IPs
   - Device tree support for the simple card, max98090 and cs42l52
   - Conversion of the Samsung drivers to native dmaengine, making them
     multiplatform compatible and hopefully helping keep them more
     modern and up to date.
   - More regmap conversions, including a very welcome one for twl6040
     from Peter Ujfalusi
   - A big overhaul of the DaVinci drivers also from Peter Ujfalusi
   - Lots of DMA updates from Lars-Peter
   - Improvements to the constraints handling code from Lars-Peter
   - A very helpful conversion of the TWL4030 driver to regmap from Peter
   - A new driver for the Freescale ESAI controller from Nicolin Chen
   - Conversion of some of the drivers to use params_width()
   - Extensions to DPCM for use with compressed audio from Liam"

* tag 'sound-3.14-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (396 commits)
  ASoC: dapm: Fix double prefix addition
  ASoC: compress: Add suport for DPCM into compressed audio
  ASoC: DPCM: make some DPCM API calls non static for compressed usage
  ASoC: core: Fix possible NULL pointer dereference of pcm->config
  ALSA: hda - add headset mic detect quirks for some Dell machines
  ASoC: tlv320aic32x4: Fix regmap range_min
  ASoC: core: Return -ENOTSUPP from set_sysclk() if no operation provided
  ASoC: dapm: Change prototype of soc_widget_read
  ASoC: samsung: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag
  ASoC: axi-{spdif,i2s}: Remove SND_DMAENGINE_PCM_FLAG_NO_RESIDUE flag
  ASoC: generic-dmaengine-pcm: Check DMA residue granularity
  ASoC: generic-dmaengine-pcm: Check NO_RESIDUE flag at runtime
  dma: pl330: Set residue_granularity
  dma: Indicate residue granularity in dma_slave_caps
  ASoC: simple-card: fix one bug to writing to the platform data
  ASoC: pcm: Use snd_pcm_rate_mask_intersect() helper
  ALSA: Add helper function for intersecting two rate masks
  ASoC: s6000: Don't mix SNDRV_PCM_RATE_CONTINUOUS with specific rates
  ASoC: fsl: Don't mix SNDRV_PCM_RATE_CONTINUOUS with specific rates
  ASoC: pcm: Properly initialize hw->rate_max
  ...

10 years agoMerge tag 'pinctrl-v3.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Tue, 21 Jan 2014 18:14:10 +0000 (10:14 -0800)]
Merge tag 'pinctrl-v3.14-1' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull bulk pin control changes from Linus Walleij:
 "This has been queued and tested for a while.  Lots of action here,
  like in the GPIO tree, embedded stuff like this is really hot now it
  seems.  Details in the signed tag.  I'm especially happy about the
  Qualcomm driver as it is used in such a huge subset of mobile handsets
  out there, and these platforms in general need better upstream support

   - New driver for the Qualcomm TLMM pin controller and its msm8x74
     subdriver.

   - New driver for the Broadcom Capri BCM281xx SoC.

   - New subdriver for the imx25 pin controller.

   - New subdriver for the Tegra124 pin controller.

   - Lock GPIO lines as IRQs for select combined pin control and GPIO
     drivers for baytrail and sirf.

   - Some semi-big refactorings and extenstions to the sirf driver.

   - Lots of patching, cleanup and fixing in the Renesas "PFC" driver
     and associated subdrivers as usual.  It is settling down a little
     bit now it seems.

   - Minor fixes and incremental updates here and there as usual"

* tag 'pinctrl-v3.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl: (72 commits)
  pinctrl: sunxi: Honor GPIO output initial vaules
  pinctrl: capri: add dependency on OF
  ARM: bcm11351: Enable pinctrl for Broadcom Capri SoCs
  ARM: pinctrl: Add Broadcom Capri pinctrl driver
  pinctrl: Add pinctrl binding for Broadcom Capri SoCs
  pinctrl: Add void * to pinctrl_pin_desc
  pinctrl: st: Fix a typo in probe
  pinctrl: Fix some typos and grammar issues in the documentation
  pinctrl: sirf: lock IRQs when starting them
  pinctrl: sirf: put gpio interrupt pin into input status automatically
  pinctrl: sirf: use only one irq_domain for the whole device node
  pinctrl: single: fix infinite loop caused by bad mask
  pinctrl: single: fix pcs_disable with bits_per_mux
  pinctrl: single: fix DT bindings documentation
  pinctrl: as3722: Set pin to output mode for some function
  pinctrl: sirf: add pin group for USP0 with only RX or TX frame sync
  pinctrl: sirf: fix the pins of sdmmc5 connected with TriG
  pinctrl: sirf: add lost usp1_uart_nostreamctrl group for atlas6
  pinctrl: sunxi: Add Allwinner A20 clock output pin functions
  pinctrl/lantiq: fix typo
  ...

10 years agoMerge tag 'gpio-v3.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Tue, 21 Jan 2014 17:40:46 +0000 (09:40 -0800)]
Merge tag 'gpio-v3.14-1' of git://git./linux/kernel/git/linusw/linux-gpio

Pull GPIO tree bulk changes from Linus Walleij:
 "A big set this merge window, as we have much going on in this
  subsystem.  The changes to other subsystems (notably a slew of ARM
  machines as I am doing away with their custom APIs) have all been
  ACKed to the extent possible.

  Major changes this time:

   - Some core improvements and cleanups to the new GPIO descriptor API.
     This seems to be working now so we can start the exodus to this
     API, moving gradually away from the global GPIO numberspace.

   - Incremental improvements to the ACPI GPIO core, and move the few
     GPIO ACPI clients we have to the GPIO descriptor API right *now*
     before we go any further.  We actually managed to contain this
     *before* we started to litter the kernel with yet another hackish
     global numberspace for the ACPI GPIOs, which is a big win.

   - The RFkill GPIO driver and all platforms using it have been
     migrated to use the GPIO descriptors rather than fixed number
     assignments.  Tegra machine has been migrated as part of this.

   - New drivers for MOXA ART, Xtensa GPIO32 and SMSC SCH311x.  Those
     should be really good examples of how I expect a nice GPIO driver
     to look these days.

   - Do away with custom GPIO implementations on a major part of the ARM
     machines: ks8695, lpc32xx, mv78xx0.  Make a first step towards the
     same in the horribly convoluted Samsung S3C include forest.  We
     expect to continue to clean this up as we move forward.

   - Flag GPIO lines used for IRQ on adnp, bcm-kona, em, intel-mid and
     lynxpoint.

     This makes the GPIOlib core aware that a certain GPIO line is used
     for IRQs and can then enforce some semantics such as disallowing a
     GPIO line marked as in use for IRQ to be switched to output mode.

   - Drop all use of irq_set_chip_and_handler_name().  The name provided
     in these cases were just unhelpful tags like "mux" or "demux".

   - Extend the MCP23s08 driver to handle interrupts.

   - Minor incremental improvements for rcar, lynxpoint, em 74x164 and
     msm drivers.

   - Some non-urgent bug fixes here and there, duplicate #includes and
     that usual kind of cleanups"

Fix up broken Kconfig file manually to make this all compile.

* tag 'gpio-v3.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio: (71 commits)
  gpio: mcp23s08: fix casting caused build warning
  gpio: mcp23s08: depend on OF_GPIO
  gpio: mcp23s08: Add irq functionality for i2c chips
  ARM: S5P[v210|c100|64x0]: Fix build error
  gpio: pxa: clamp gpio get value to [0,1]
  ARM: s3c24xx: explicit dependency on <plat/gpio-cfg.h>
  ARM: S3C[24|64]xx: move includes back under <mach/> scope
  Documentation / ACPI: update to GPIO descriptor API
  gpio / ACPI: get rid of acpi_gpio.h
  gpio / ACPI: register to ACPI events automatically
  mmc: sdhci-acpi: convert to use GPIO descriptor API
  ARM: s3c24xx: fix build error
  gpio: f7188x: set can_sleep attribute
  gpio: samsung: Update documentation
  gpio: samsung: Remove hardware.h inclusion
  gpio: xtensa: depend on HAVE_XTENSA_GPIO32
  gpio: clps711x: Enable driver compilation with COMPILE_TEST
  gpio: clps711x: Use of_match_ptr()
  net: rfkill: gpio: convert to descriptor-based GPIO interface
  leds: s3c24xx: Fix build failure
  ...

10 years agoMerge branch 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Tue, 21 Jan 2014 17:39:37 +0000 (09:39 -0800)]
Merge branch 'i2c/for-next' of git://git./linux/kernel/git/wsa/linux

Pull i2c updates from Wolfram Sang:
 "For 3.14, the I2C subsystem has the following to offer:

   - new drivers for Renesas RIIC and RobotFuzz OSIF
   - driver cleanups & improvements & bugfixes

  Pretty standard stuff this time, I'd say.  There is more complex stuff
  coming up, but I didn't have the bandwidth between the years to pull
  it in for this release.  Sadly"

* 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux: (26 commits)
  i2c: s3c2410: fix quirk usage for 64-bit
  i2c: pnx: Use devm_*() functions
  i2c: at91: add a new compatibility string for the at91sam9261
  i2c-ismt: support I2C_SMBUS_I2C_BLOCK_DATA transaction type
  i2c: Add bus driver for for OSIF USB i2c device.
  i2c: i2c-tiny-usb: Remove RobotFuzz USB vendor:product ID
  i2c: designware: remove HAVE_CLK build dependecy
  Documentation: i2c: Remove obsolete example
  i2c: nomadik: remove platform data header
  i2c: nomadik: auto-calculate slave setup time
  i2c: viperboard: remove superfluous assignment
  i2c: xilinx: Use devm_* functions
  i2c: xilinx: Do not enable irq before irq handler
  i2c: xilinx: Fix i2c checkpatch warnings
  i2c: at91: document clock properties
  i2c: isch: Use devm_request_region()
  i2c: viperboard: Use devm_kzalloc() functions
  i2c: imx: propagate irq error code in probe
  i2c: s3c2410: dont need CPU_FREQ transitions for exynos series
  i2c: s3c2410: Add polling mode support
  ...

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris...
Linus Torvalds [Tue, 21 Jan 2014 17:06:02 +0000 (09:06 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jmorris/linux-security

Pull security layer updates from James Morris:
 "Changes for this kernel include maintenance updates for Smack, SELinux
  (and several networking fixes), IMA and TPM"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security: (39 commits)
  SELinux: Fix memory leak upon loading policy
  tpm/tpm-sysfs: active_show() can be static
  tpm: tpm_tis: Fix compile problems with CONFIG_PM_SLEEP/CONFIG_PNP
  tpm: Make tpm-dev allocate a per-file structure
  tpm: Use the ops structure instead of a copy in tpm_vendor_specific
  tpm: Create a tpm_class_ops structure and use it in the drivers
  tpm: Pull all driver sysfs code into tpm-sysfs.c
  tpm: Move sysfs functions from tpm-interface to tpm-sysfs
  tpm: Pull everything related to /dev/tpmX into tpm-dev.c
  char: tpm: nuvoton: remove unused variable
  tpm: MAINTAINERS: Cleanup TPM Maintainers file
  tpm/tpm_i2c_atmel: fix coccinelle warnings
  tpm/tpm_ibmvtpm: fix unreachable code warning (smatch warning)
  tpm/tpm_i2c_stm_st33: Check return code of get_burstcount
  tpm/tpm_ppi: Check return value of acpi_get_name
  tpm/tpm_ppi: Do not compare strcmp(a,b) == -1
  ima: remove unneeded size_limit argument from ima_eventdigest_init_common()
  ima: update IMA-templates.txt documentation
  ima: pass HASH_ALGO__LAST as hash algo in ima_eventdigest_init()
  ima: change the default hash algorithm to SHA1 in ima_eventdigest_ng_init()
  ...

10 years agomfd: wm5110: Add register patch for rev D chip
Charles Keepax [Thu, 9 Jan 2014 11:53:54 +0000 (11:53 +0000)]
mfd: wm5110: Add register patch for rev D chip

Evaluation of revision D of WM5110 suggests updates to the register
patch for optimal performance. For the sake of clarity rev C of the chip
does not require a register patch.

Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: omap-usb-tll: Don't hold lock during pm_runtime_get/put_sync()
Roger Quadros [Wed, 8 Jan 2014 07:15:33 +0000 (12:45 +0530)]
mfd: omap-usb-tll: Don't hold lock during pm_runtime_get/put_sync()

pm_runtime_get/put_sync() can sleep so don't hold spinlock while
calling them.

This patch prevents a BUG() during system suspend when
CONFIG_DEBUG_ATOMIC_SLEEP is enabled.

Bug is present in Kernel versions v3.9 onwards.

Reported-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Tested-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agogpio: lp3943: Remove redundant of_match_ptr helper
Sachin Kamat [Tue, 7 Jan 2014 06:27:18 +0000 (11:57 +0530)]
gpio: lp3943: Remove redundant of_match_ptr helper

'lp3943_gpio_of_match' is always compiled in. Hence the
helper macro is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: sta2x11-mfd: Use named constants for pci_power_t values
Julia Lawall [Thu, 2 Jan 2014 23:40:29 +0000 (00:40 +0100)]
mfd: sta2x11-mfd: Use named constants for pci_power_t values

If nothing more than to improve code readability.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression pdev;
@@

pci_set_power_state(pdev,
- 0
+ PCI_D0
 )
// </smpl>

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agoDocumentation: mfd: Fix LDO index in s2mps11.txt
Sachin Kamat [Tue, 24 Dec 2013 09:23:25 +0000 (14:53 +0530)]
Documentation: mfd: Fix LDO index in s2mps11.txt

LDO indices start from 1. Fix the example appropriately.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: Cleanup mfd-mcp-sa11x0.h header
Sachin Kamat [Mon, 30 Dec 2013 06:24:56 +0000 (11:54 +0530)]
mfd: Cleanup mfd-mcp-sa11x0.h header

Commit a1fd844c6e62 ("ARM: sa1100: move platform_data definitions")
moved the file to the current location but forgot to remove the pointer
to its previous location. Clean it up. While at it also change the header
file protection macros appropriately.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: max8997: Use "IS_ENABLED(CONFIG_OF)" for DT code.
Manish Badarkhe [Sun, 22 Dec 2013 17:48:49 +0000 (23:18 +0530)]
mfd: max8997: Use "IS_ENABLED(CONFIG_OF)" for DT code.

Instead of "#if define CONFIG_OF" use "IS_ENABLED(CONFIG_OF)"
option for DT code to avoid if-deffery in code.

Signed-off-by: Manish Badarkhe <badarkhe.manish@gmail.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: twl6030: Fix endianness problem in IRQ handler
Danke Xie [Mon, 23 Dec 2013 17:11:46 +0000 (19:11 +0200)]
mfd: twl6030: Fix endianness problem in IRQ handler

The current TWL 6030 IRQ handler assumes little endianness.
This change makes it endian-neutral.

Signed-off-by: Danke Xie <d.xie@sta.samsung.com>
Signed-off-by: Taras Kondratiuk <taras.kondratiuk@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: sec-core: Add cells for S5M8767-clocks
Tushar Behera [Thu, 26 Dec 2013 10:19:00 +0000 (15:49 +0530)]
mfd: sec-core: Add cells for S5M8767-clocks

S5M8767 chip has 3 crystal oscillators running at 32KHz. These are
supported by s2mps11-clk driver.

Signed-off-by: Tushar Behera <tushar.behera@linaro.org>
Reviewed-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: max14577: Remove redundant of_match_ptr helper
Sachin Kamat [Sat, 21 Dec 2013 10:20:16 +0000 (15:50 +0530)]
mfd: max14577: Remove redundant of_match_ptr helper

'max14577_dt_match' is always compiled in. Hence the helper macro
is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: twl6040: Fix sparse non static symbol warning
Wei Yongjun [Sat, 21 Dec 2013 02:14:13 +0000 (10:14 +0800)]
mfd: twl6040: Fix sparse non static symbol warning

Fixes the following sparse warning:

drivers/mfd/twl6040.c:89:20: warning:
 symbol 'twl6040_patch' was not declared. Should it be static?

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: Revert "mfd: Always assign of_node in mfd_add_device()"
Lee Jones [Fri, 20 Dec 2013 14:21:16 +0000 (14:21 +0000)]
mfd: Revert "mfd: Always assign of_node in mfd_add_device()"

This reverts commit 68044bee13770918e0b28dd44aa98c889ec7558f.

We've had confirmed reports of this patch causing unforeseen issues
with existing MFD users. It has been agreed by the original author
and myself that reversion is the best solution.

Acked-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: rtsx: Fix sparse non static symbol warning
Wei Yongjun [Fri, 20 Dec 2013 02:50:30 +0000 (10:50 +0800)]
mfd: rtsx: Fix sparse non static symbol warning

Fixes the following sparse warning:

drivers/mfd/rtl8411.c:473:6: warning:
 symbol 'rtl8411_init_common_params' was not declared. Should it be static?

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: max77693: Set proper maximum register for MUIC regmap
Krzysztof Kozlowski [Fri, 20 Dec 2013 09:35:08 +0000 (10:35 +0100)]
mfd: max77693: Set proper maximum register for MUIC regmap

The MUIC block in max77693 has different I2C address than PMIC. The
driver allocated two regmaps: for PMIC and MUIC. However it used the
same regmap_config (with max_register field) for both regmaps. Actual
maximum address of register for MUIC is different than for PMIC.

Define another regmap_config for MUIC with proper max_register value.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: max77686: Fix regmap resource leak on driver remove
Krzysztof Kozlowski [Fri, 20 Dec 2013 09:35:07 +0000 (10:35 +0100)]
mfd: max77686: Fix regmap resource leak on driver remove

The regmap used by max77686 MFD driver was not freed with regmap_exit()
on driver exit. This lead to leak of resources.

Replace regmap_init_i2c() call in driver probe with initialization of
managed register map so the regmap will be properly freed by the device
management code.

Cc: stable@vger.kernel.org
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: Represent correct filenames in file headers
Laszlo Papp [Thu, 19 Dec 2013 13:08:43 +0000 (13:08 +0000)]
mfd: Represent correct filenames in file headers

The original author(s) probably copy/pasted these headers from the
existing public header files.

Signed-off-by: Laszlo Papp <lpapp@kde.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: rtsx: Add support for card reader rtl8402
Micky Ching [Wed, 18 Dec 2013 02:03:13 +0000 (10:03 +0800)]
mfd: rtsx: Add support for card reader rtl8402

rtl8402 is much like rtl8411, so just add it to rtl8411.c

Signed-off-by: Micky Ching <micky_ching@realsil.com.cn>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: rtsx: Add set pull control macro and simplify rtl8411
Micky Ching [Wed, 18 Dec 2013 02:03:12 +0000 (10:03 +0800)]
mfd: rtsx: Add set pull control macro and simplify rtl8411

Add set pull control macro to reduce code for setting pull control, and
use a common init function to reduce code for rtl8411.c. So this patch
is used to just simplify code.

Signed-off-by: Micky Ching <micky_ching@realsil.com.cn>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: max8997: Enforce mfd_add_devices() return value check
Laszlo Papp [Tue, 17 Dec 2013 12:49:51 +0000 (12:49 +0000)]
mfd: max8997: Enforce mfd_add_devices() return value check

The original author provided a random return value check which is
redundant and seemingly floating. This patch not only relocates
the check so it is more clearly associated with the invokation of
mfd_add_devices(), but provides a store for the error value. We
also print a meaningful message on error before returning.

Signed-off-by: Laszlo Papp <lpapp@kde.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: mc13xxx: Simplify probe() & remove()
Alexander Shiyan [Sat, 14 Dec 2013 13:03:12 +0000 (17:03 +0400)]
mfd: mc13xxx: Simplify probe() & remove()

This patch simplifies probe() and remove() functions by moving
some initialisation code out from the I2C/SPI init() and exit()
functions and into the core driver.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: mc13xxx: Remove useless symbol MFD_MC13783
Alexander Shiyan [Sat, 14 Dec 2013 13:03:10 +0000 (17:03 +0400)]
mfd: mc13xxx: Remove useless symbol MFD_MC13783

Symbol MFD_MC13783 always selected by MFD_MC13XXX, so no need
to keep additional symbol.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: mc13xxx: Remove duplicate mc13xxx_get_flags() declaration
Alexander Shiyan [Sat, 14 Dec 2013 13:03:11 +0000 (17:03 +0400)]
mfd: mc13xxx: Remove duplicate mc13xxx_get_flags() declaration

mc13xxx_get_flags() declaration given twice.
This patch removes this duplicate.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: ssbi: Mark match table const
Stephen Boyd [Tue, 10 Dec 2013 23:35:19 +0000 (15:35 -0800)]
mfd: ssbi: Mark match table const

This is a read-only data structure.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: ssbi: Constify buffer in ssbi_write
Stephen Boyd [Tue, 10 Dec 2013 23:35:17 +0000 (15:35 -0800)]
mfd: ssbi: Constify buffer in ssbi_write

In preparation for passing a const pointer directly to
ssbi_write() from the regmap APIs.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: ssbi: Remove platform data structs and hide ssbi type enum
Stephen Boyd [Tue, 10 Dec 2013 23:35:16 +0000 (15:35 -0800)]
mfd: ssbi: Remove platform data structs and hide ssbi type enum

The ssbi driver assumes that the device is DT based. Remove the
platform data structs that will never be used and hide the enum
in the only C file that uses it.

Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: Always assign of_node in mfd_add_device()
Stephen Warren [Mon, 9 Dec 2013 22:06:47 +0000 (15:06 -0700)]
mfd: Always assign of_node in mfd_add_device()

mfd_add_device() assigns .of_node in the device objects it creates only
if the mfd_cell for the device has the .of_compatible field set and the
DT node for the top-level MFD device contains a child whose compatible
property matches the cell's .of_compatible field.

This leaves .of_node unset in many cases. When this happens, entries in
the DT /aliases property which refer to the top-level MFD DT node will
never match the MFD child devices, hence causing the requested alias not
to be honored.

Solve this by setting each MFD child device's .of_node equal to the top-
level MFD device's .of_node field in the cases where it would otherwise
remain unset.

The first use-case for this will be aliases for the TPS6586x's RTC
device.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: cros ec: spi: Add delay for raising CS
Rhyland Klein [Mon, 9 Dec 2013 11:36:09 +0000 (12:36 +0100)]
mfd: cros ec: spi: Add delay for raising CS

The EC has specific timing it requires. Add support for an optional delay
after raising CS to fix timing issues. This is configurable based on
a DT property "google,cros-ec-spi-msg-delay".

If this property isn't set, then no delay will be added. However, if set
it will cause a delay equal to the value passed to it to be inserted at
the end of a transaction.

Signed-off-by: Rhyland Klein <rklein@nvidia.com>
Reviewed-by: Bernie Thompson <bhthompson@chromium.org>
Reviewed-by: Andrew Bresticker <abrestic@chromium.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: cros ec: spi: Use 0 instead of '\0' consistently
Thierry Reding [Mon, 9 Dec 2013 11:36:10 +0000 (12:36 +0100)]
mfd: cros ec: spi: Use 0 instead of '\0' consistently

memset() was being called with the second parameter set to '\0', which
is equivalent but longer than the more canonical 0. Update the code to
use the latter variant consistently across the driver.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: cros ec: spi: Depend on OF
Thierry Reding [Mon, 9 Dec 2013 10:05:39 +0000 (11:05 +0100)]
mfd: cros ec: spi: Depend on OF

The driver is not used on any non-DT platform, so it can depend on the
OF Kconfig symbol to make that explicit. This is in preparation of a
subsequent patch which parses some parameters from the device tree and
would otherwise have to conditionalize that code.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: cros ec: spi: Use correct module license
Thierry Reding [Mon, 9 Dec 2013 10:05:38 +0000 (11:05 +0100)]
mfd: cros ec: spi: Use correct module license

According to the header comment in the source file the driver is
licensed under GPL v2, so update MODULE_LICENSE() to match that.

Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: sec-core: Fix sparse NULL pointer warning
Wei Yongjun [Sat, 7 Dec 2013 06:02:15 +0000 (14:02 +0800)]
mfd: sec-core: Fix sparse NULL pointer warning

Fixes the following sparse warning:

drivers/mfd/sec-core.c:202:16: warning: Using plain integer as NULL pointer

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
10 years agomfd: sta2x11-mfd: Fix return value check in sta2x11_mfd_platform_probe()
Wei Yongjun [Sat, 7 Dec 2013 06:03:00 +0000 (14:03 +0800)]
mfd: sta2x11-mfd: Fix return value check in sta2x11_mfd_platform_probe()

In case of error, the function devm_regmap_init_mmio() returns
ERR_PTR() and never returns NULL. The NULL test in the return
value check should be replaced with IS_ERR().

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Lee Jones <lee.jones@linaro.org>