| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The "strictlimit" feature was introduced to enforce per-bdi dirty limits
for FUSE which sets bdi max_ratio to 1% by default:
http://article.gmane.org/gmane.linux.kernel.mm/105809
However the feature can be useful for other relatively slow or untrusted
BDIs like USB flash drives and DVD+RW. The patch adds a knob to enable
the feature:
echo 1 > /sys/class/bdi/X:Y/strictlimit
Being enabled, the feature enforces bdi max_ratio limit even if global
(10%) dirty limit is not reached. Of course, the effect is not visible
until /sys/class/bdi/X:Y/max_ratio is decreased to some reasonable value.
Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com>
Cc: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Cc: Theodore Ts'o <tytso@mit.edu>
Cc: "Artem S. Tashkinov" <t.artem@lycos.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Jan Kara <jack@suse.cz>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Add my copyright to the zsmalloc source code which I maintain.
Signed-off-by: Minchan Kim <minchan@kernel.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
This patch moves zsmalloc under mm directory.
Before that, description will explain why we have needed custom allocator.
Zsmalloc is a new slab-based memory allocator for storing compressed
pages. It is designed for low fragmentation and high allocation success
rate on large object, but <= PAGE_SIZE allocations.
zsmalloc differs from the kernel slab allocator in two primary ways to
achieve these design goals.
zsmalloc never requires high order page allocations to back slabs, or
"size classes" in zsmalloc terms. Instead it allows multiple single-order
pages to be stitched together into a "zspage" which backs the slab. This
allows for higher allocation success rate under memory pressure.
Also, zsmalloc allows objects to span page boundaries within the zspage.
This allows for lower fragmentation than could be had with the kernel slab
allocator for objects between PAGE_SIZE/2 and PAGE_SIZE. With the kernel
slab allocator, if a page compresses to 60% of it original size, the
memory savings gained through compression is lost in fragmentation because
another object of the same size can't be stored in the leftover space.
This ability to span pages results in zsmalloc allocations not being
directly addressable by the user. The user is given an non-dereferencable
handle in response to an allocation request. That handle must be mapped,
using zs_map_object(), which returns a pointer to the mapped region that
can be used. The mapping is necessary since the object data may reside in
two different noncontigious pages.
The zsmalloc fulfills the allocation needs for zram perfectly
[sjenning@linux.vnet.ibm.com: borrow Seth's quote]
Signed-off-by: Minchan Kim <minchan@kernel.org>
Acked-by: Nitin Gupta <ngupta@vflare.org>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Luigi Semenzato <semenzato@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Seth Jennings <sjenning@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
7851a45cd3 ("mm: numa: Copy cpupid on page migration") copies over the
cpupid at page migration time. it is unnecessary to set it again in
alloc_misplaced_dst_page().
Signed-off-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
7851a45cd3 (mm: numa: Copy cpupid on page migration) copies over the
cpupid at page migration time. It is unnecessary to set it again in
migrate_misplaced_transhuge_page().
Signed-off-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Conflicts:
drivers/misc/mei/hbm.c
|
|
|
|
Conflicts:
fs/btrfs/extent_io.c
fs/f2fs/data.c
fs/f2fs/segment.c
include/trace/events/f2fs.h
|
|
|
|
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Shaohua Li <shli@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
This is a patch to improve swap readahead algorithm. It's from Hugh and I
slightly changed it.
Hugh's original changelog:
swapin readahead does a blind readahead, whether or not the swapin
is sequential. This may be ok on harddisk, because large reads have
relatively small costs, and if the readahead pages are unneeded they
can be reclaimed easily - though, what if their allocation forced
reclaim of useful pages? But on SSD devices large reads are more
expensive than small ones: if the readahead pages are unneeded,
reading them in caused significant overhead.
This patch adds very simplistic random read detection. Stealing
the PageReadahead technique from Konstantin Khlebnikov's patch,
avoiding the vma/anon_vma sophistications of Shaohua Li's patch,
swapin_nr_pages() simply looks at readahead's current success
rate, and narrows or widens its readahead window accordingly.
There is little science to its heuristic: it's about as stupid
as can be whilst remaining effective.
The table below shows elapsed times (in centiseconds) when running
a single repetitive swapping load across a 1000MB mapping in 900MB
ram with 1GB swap (the harddisk tests had taken painfully too long
when I used mem=500M, but SSD shows similar results for that).
Vanilla is the 3.6-rc7 kernel on which I started; Shaohua denotes
his Sep 3 patch in mmotm and linux-next; HughOld denotes my Oct 1
patch which Shaohua showed to be defective; HughNew this Nov 14
patch, with page_cluster as usual at default of 3 (8-page reads);
HughPC4 this same patch with page_cluster 4 (16-page reads);
HughPC0 with page_cluster 0 (1-page reads: no readahead).
HDD for swapping to harddisk, SSD for swapping to VertexII SSD.
Seq for sequential access to the mapping, cycling five times around;
Rand for the same number of random touches. Anon for a MAP_PRIVATE
anon mapping; Shmem for a MAP_SHARED anon mapping, equivalent to tmpfs.
One weakness of Shaohua's vma/anon_vma approach was that it did
not optimize Shmem: seen below. Konstantin's approach was perhaps
mistuned, 50% slower on Seq: did not compete and is not shown below.
HDD Vanilla Shaohua HughOld HughNew HughPC4 HughPC0
Seq Anon 73921 76210 75611 76904 78191 121542
Seq Shmem 73601 73176 73855 72947 74543 118322
Rand Anon 895392 831243 871569 845197 846496 841680
Rand Shmem 1058375 1053486 827935 764955 764376 756489
SSD Vanilla Shaohua HughOld HughNew HughPC4 HughPC0
Seq Anon 24634 24198 24673 25107 21614 70018
Seq Shmem 24959 24932 25052 25703 22030 69678
Rand Anon 43014 26146 28075 25989 26935 25901
Rand Shmem 45349 45215 28249 24268 24138 24332
These tests are, of course, two extremes of a very simple case:
under heavier mixed loads I've not yet observed any consistent
improvement or degradation, and wider testing would be welcome.
Shaohua Li:
Test shows Vanilla is slightly better in sequential workload than Hugh's patch.
I observed with Hugh's patch sometimes the readahead size is shrinked too fast
(from 8 to 1 immediately) in sequential workload if there is no hit. And in
such case, continuing doing readahead is good actually.
I don't prepare a sophisticated algorithm for the sequential workload because
so far we can't guarantee sequential accessed pages are swap out sequentially.
So I slightly change Hugh's heuristic - don't shrink readahead size too fast.
Here is my test result (unit second, 3 runs average):
Vanilla Hugh New
Seq 356 370 360
Random 4525 2447 2444
Attached graph is the swapin/swapout throughput I collected with 'vmstat 2'.
The first part is running a random workload (till around 1200 of the x-axis)
and the second part is running a sequential workload. swapin and swapout
throughput are almost identical in steady state in both workloads. These are
expected behavior. while in Vanilla, swapin is much bigger than swapout
especially in random workload (because wrong readahead).
Original patches by: Shaohua Li and Konstantin Khlebnikov.
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Shaohua Li <shli@fusionio.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Konstantin Khlebnikov <khlebnikov@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Fix race between swapoff and swapon resulting in setting blocksize of
PAGE_SIZE for block devices during swapoff.
The swapon modifies swap_info->old_block_size before acquiring
swapon_mutex. It reads block_size of bdev, stores it under
swap_info->old_block_size and sets new block_size to PAGE_SIZE.
On the other hand the swapoff sets the device's block_size to
old_block_size after releasing swapon_mutex.
This patch locks the swapon_mutex much earlier during swapon. It also
releases the swapon_mutex later during swapoff.
The effect of race can be triggered by following scenario:
- One block swap device with block size of 512
- thread 1: Swapon is called, swap is activated,
p->old_block_size = block_size(p->bdev); /512/
block_size(p->bdev) = PAGE_SIZE;
Thread ends.
- thread 2: Swapoff is called and it goes just after releasing the
swapon_mutex. The swap is now fully disabled except of setting the
block size to old value. The p->bdev->block_size is still equal to
PAGE_SIZE.
- thread 3: New swapon is called. This swap is disabled so without
acquiring the swapon_mutex:
- p->old_block_size = block_size(p->bdev); /PAGE_SIZE (!!!)/
- block_size(p->bdev) = PAGE_SIZE;
Swap is activated and thread ends.
- thread 2: resumes work and sets blocksize to old value:
- set_blocksize(bdev, p->old_block_size)
But now the p->old_block_size is equal to PAGE_SIZE.
The patch swap-fix-set_blocksize-race-during-swapon-swapoff does not fix
this particular issue. It reduces the possibility of races as the swapon
must overwrite p->old_block_size before acquiring swapon_mutex in swapoff.
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Cc: Weijie Yang <weijie.yang.kh@gmail.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Shaohua Li <shli@fusionio.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Eric has reported that he can see task(s) stuck in memcg OOM handler
regularly. The only way out is to
echo 0 > $GROUP/memory.oom_controll
His usecase is:
- Setup a hierarchy with memory and the freezer (disable kernel oom and
have a process watch for oom).
- In that memory cgroup add a process with one thread per cpu.
- In one thread slowly allocate once per second I think it is 16M of ram
and mlock and dirty it (just to force the pages into ram and stay
there).
- When oom is achieved loop:
* attempt to freeze all of the tasks.
* if frozen send every task SIGKILL, unfreeze, remove the directory in
cgroupfs.
Eric has then pinpointed the issue to be memcg specific.
All tasks are sitting on the memcg_oom_waitq when memcg oom is disabled.
Those that have received fatal signal will bypass the charge and should
continue on their way out. The tricky part is that the exit path might
trigger a page fault (e.g. exit_robust_list), thus the memcg charge,
while its memcg is still under OOM because nobody has released any charges
yet.
Unlike with the in-kernel OOM handler the exiting task doesn't get
TIF_MEMDIE set so it doesn't shortcut futher charges of the killed task
and falls to the memcg OOM again without any way out of it as there are no
fatal signals pending anymore.
This patch fixes the issue by checking PF_EXITING early in
__mem_cgroup_try_charge and bypass the charge same as if it had fatal
signal pending or TIF_MEMDIE set.
Normally exiting tasks (aka not killed) will bypass the charge now but
this should be OK as the task is leaving and will release memory and
increasing the memory pressure just to release it in a moment seems
dubious wasting of cycles. Besides that charges after exit_signals should
be rare.
Reported-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
If echo -1 > /proc/vm/sys/min_free_kbytes, the system will hang. Changing
proc_dointvec() to proc_dointvec_minmax() in the
min_free_kbytes_sysctl_handler() can prevent this to happen.
mhocko said:
: You can still do echo $BIG_VALUE > /proc/vm/sys/min_free_kbytes and make
: your machine unusable but I agree that proc_dointvec_minmax is more
: suitable here as we already have:
:
: .proc_handler = min_free_kbytes_sysctl_handler,
: .extra1 = &zero,
:
: It used to work properly but then 6fce56ec91b5 ("sysctl: Remove references
: to ctl_name and strategy from the generic sysctl table") has removed
: sysctl_intvec strategy and so extra1 is ignored.
Signed-off-by: Han Pingtian <hanpt@linux.vnet.ibm.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
11c731e81bb0 ("mm/mempolicy: fix !vma in new_vma_page()") has removed
BUG_ON(!vma) from new_vma_page which is partially correct because
page_address_in_vma will return EFAULT for non-linear mappings and at
least shared shmem might be mapped this way.
The patch also tried to prevent NULL ptr for hugetlb pages which is not
correct AFAICS because hugetlb pages cannot be mapped as VM_NONLINEAR and
other conditions in page_address_in_vma seem to be legit and catch real
bugs.
This patch restores BUG_ON for PageHuge to catch potential issues when the
to-be-migrated page is not setup properly.
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Reviewed-by: Bob Liu <bob.liu@oracle.com>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
After thp split in hwpoison_user_mappings(), we hold page lock on the raw
error page only between try_to_unmap, hence we are in danger of race condition.
I found in the RHEL7 MCE-relay testing that we have "bad page" error
when a memory error happens on a thp tail page used by qemu-kvm:
Triggering MCE exception on CPU 10
mce: [Hardware Error]: Machine check events logged
MCE exception done on CPU 10
MCE 0x38c535: Killing qemu-kvm:8418 due to hardware memory corruption
MCE 0x38c535: dirty LRU page recovery: Recovered
qemu-kvm[8418]: segfault at 20 ip 00007ffb0f0f229a sp 00007fffd6bc5240 error 4 in qemu-kvm[7ffb0ef14000+420000]
BUG: Bad page state in process qemu-kvm pfn:38c400
page:ffffea000e310000 count:0 mapcount:0 mapping: (null) index:0x7ffae3c00
page flags: 0x2fffff0008001d(locked|referenced|uptodate|dirty|swapbacked)
Modules linked in: hwpoison_inject mce_inject vhost_net macvtap macvlan ...
CPU: 0 PID: 8418 Comm: qemu-kvm Tainted: G M -------------- 3.10.0-54.0.1.el7.mce_test_fixed.x86_64 #1
Hardware name: NEC NEC Express5800/R120b-1 [N8100-1719F]/MS-91E7-001, BIOS 4.6.3C19 02/10/2011
000fffff00000000 ffff8802fc9239a0 ffffffff815b4cc0 ffff8802fc9239b8
ffffffff815b072e 0000000000000000 ffff8802fc9239f8 ffffffff8113b918
ffffea000e310000 ffffea000e310000 002fffff0008001d 0000000000000000
Call Trace:
[<ffffffff815b4cc0>] dump_stack+0x19/0x1b
[<ffffffff815b072e>] bad_page.part.59+0xcf/0xe8
[<ffffffff8113b918>] free_pages_prepare+0x148/0x160
[<ffffffff8113c231>] free_hot_cold_page+0x31/0x140
[<ffffffff8113c386>] free_hot_cold_page_list+0x46/0xa0
[<ffffffff81141361>] release_pages+0x1c1/0x200
[<ffffffff8116e47d>] free_pages_and_swap_cache+0xad/0xd0
[<ffffffff8115850c>] tlb_flush_mmu.part.46+0x4c/0x90
[<ffffffff81159045>] tlb_finish_mmu+0x55/0x60
[<ffffffff81163e6b>] exit_mmap+0xcb/0x170
[<ffffffff81055f87>] mmput+0x67/0xf0
[<ffffffffa05c7451>] vhost_dev_cleanup+0x231/0x260 [vhost_net]
[<ffffffffa05ca0df>] vhost_net_release+0x3f/0x90 [vhost_net]
[<ffffffff8119f649>] __fput+0xe9/0x270
[<ffffffff8119f8fe>] ____fput+0xe/0x10
[<ffffffff8107b754>] task_work_run+0xc4/0xe0
[<ffffffff8105e88b>] do_exit+0x2bb/0xa40
[<ffffffff8106b0cc>] ? __dequeue_signal+0x13c/0x220
[<ffffffff8105f08f>] do_group_exit+0x3f/0xa0
[<ffffffff8106dcc0>] get_signal_to_deliver+0x1d0/0x6e0
[<ffffffff81012408>] do_signal+0x48/0x5e0
[<ffffffff8106ee48>] ? do_sigaction+0x88/0x1f0
[<ffffffff81012a11>] do_notify_resume+0x71/0xc0
[<ffffffff815bc53c>] retint_signal+0x48/0x8c
The reason of this bug is that a page fault happens before unlocking
the head page at the end of memory_failure().
This strange page fault is trying to access to address 0x20 and I'm not
sure why qemu-kvm does this, but anyway as a result the SIGSEGV makes
qemu-kvm exit and on the way we catch the bad page bug/warning because
we try to free a locked page (which was the former head page.)
To fix this, this patch suggests to shift page lock from head page to
tail page just after thp split. SIGSEGV still happens, but it affects
only error affected VMs, not a whole system.
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> [3.9+] # a3e0f9e47d5ef "mm/memory-failure.c: transfer page count from head page to tail page after split thp"
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Add a working sysctl to enable/disable automatic numa memory balancing at
runtime.
This allows us to track down performance problems with this feature and is
generally a good idea.
This was possible earlier through debugfs, but only with special debugging
options set. Also fix the boot message.
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Am Wed, 8 Jan 2014 12:08:04 +0800
schrieb Jianguo Wu <wujianguo@huawei.com>:
> For some archs, like arm64, would use memblock.memory after system
> booting, so we can not simply released to the buddy allocator, maybe
> need !defined(CONFIG_ARCH_DISCARD_MEMBLOCK).
Signed-off-by: Philipp Hachtmann <phacht@linux.vnet.ibm.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When calling free_all_bootmem() the free areas under memblock's control
are released to the buddy allocator. Additionally the reserved list is
freed if it was reallocated by memblock. The same should apply for the
memory list.
Signed-off-by: Philipp Hachtmann <phacht@linux.vnet.ibm.com>
Reviewed-by: Tejun Heo <tj@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Toshi Kani <toshi.kani@hp.com>
Cc: Jianguo Wu <wujianguo@huawei.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When memblock_reserve() fails because memblock.reserved.regions cannot be
resized, the caller (e.g. alloc_bootmem()) is not informed of the failed
allocation. Therefore alloc_bootmem() silently returns the same pointer
again and again.
This patch adds a check for the return value of memblock_reserve() in
__alloc_memory_core().
Signed-off-by: Philipp Hachtmann <phacht@linux.vnet.ibm.com>
Reviewed-by: Tejun Heo <tj@kernel.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Tang Chen <tangchen@cn.fujitsu.com>
Cc: Toshi Kani <toshi.kani@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently we take both the memcg_create_mutex and the set_limit_mutex when
we enable kmem accounting for a memory cgroup, which makes kmem activation
events serialize with both memcg creations and other memcg limit updates
(memory.limit, memory.memsw.limit). However, there is no point in such
strict synchronization rules there.
First, the set_limit_mutex was introduced to keep the memory.limit and
memory.memsw.limit values in sync. Since memory.kmem.limit can be set
independently of them, it is better to introduce a separate mutex to
synchronize against concurrent kmem limit updates.
Second, we take the memcg_create_mutex in order to make sure all children
of this memcg will be kmem-active as well. For achieving that, it is
enough to hold this mutex only while checking if memcg_has_children()
though. This guarantees that if a child is added after we checked that
the memcg has no children, the newly added cgroup will see its parent
kmem-active (of course if the latter succeeded), and call kmem activation
for itself.
This patch simplifies the locking rules of memcg_update_kmem_limit()
according to these considerations.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently we have two state bits in mem_cgroup::kmem_account_flags
regarding kmem accounting activation, ACTIVATED and ACTIVE. We start kmem
accounting only if both flags are set (memcg_can_account_kmem()), plus
throughout the code there are several places where we check only the
ACTIVE flag, but we never check the ACTIVATED flag alone. These flags are
both set from memcg_update_kmem_limit() under the set_limit_mutex, the
ACTIVE flag always being set after ACTIVATED, and they never get cleared.
That said checking if both flags are set is equivalent to checking only
for the ACTIVE flag, and since there is no ACTIVATED flag checks, we can
safely remove the ACTIVATED flag, and nothing will change.
Let's try to understand what was the reason for introducing these flags.
The purpose of the ACTIVE flag is clear - it states that kmem should be
accounting to the cgroup. The only requirement for it is that it should
be set after we have fully initialized kmem accounting bits for the cgroup
and patched all static branches relating to kmem accounting. Since we
always check if static branch is enabled before actually considering if we
should account (otherwise we wouldn't benefit from static branching), this
guarantees us that we won't skip a commit or uncharge after a charge due
to an unpatched static branch.
Now let's move on to the ACTIVATED bit. As I proved in the beginning of
this message, it is absolutely useless, and removing it will change
nothing. So what was the reason introducing it?
The ACTIVATED flag was introduced by commit a8964b9b ("memcg: use static
branches when code not in use") in order to guarantee that
static_key_slow_inc(&memcg_kmem_enabled_key) would be called only once for
each memory cgroup when its kmem accounting was activated. The point was
that at that time the memcg_update_kmem_limit() function's work-flow
looked like this:
bool must_inc_static_branch = false;
cgroup_lock();
mutex_lock(&set_limit_mutex);
if (!memcg->kmem_account_flags && val != RESOURCE_MAX) {
/* The kmem limit is set for the first time */
ret = res_counter_set_limit(&memcg->kmem, val);
memcg_kmem_set_activated(memcg);
must_inc_static_branch = true;
} else
ret = res_counter_set_limit(&memcg->kmem, val);
mutex_unlock(&set_limit_mutex);
cgroup_unlock();
if (must_inc_static_branch) {
/* We can't do this under cgroup_lock */
static_key_slow_inc(&memcg_kmem_enabled_key);
memcg_kmem_set_active(memcg);
}
So that without the ACTIVATED flag we could race with other threads trying
to set the limit and increment the static branching ref-counter more than
once. Today we call the whole memcg_update_kmem_limit() function under
the set_limit_mutex and this race is impossible.
As now we understand why the ACTIVATED bit was introduced and why we don't
need it now, and know that removing it will change nothing anyway, let's
get rid of it.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
We relocate root cache's memcg_params whenever we need to grow the
memcg_caches array to accommodate all kmem-active memory cgroups.
Currently on relocation we free the old version immediately, which can
lead to use-after-free, because the memcg_caches array is accessed
lock-free (see cache_from_memcg_idx()). This patch fixes this by making
memcg_params RCU-protected for root caches.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
There is no point in flooding logs with warnings or especially crashing
the system if we fail to create a cache for a memcg. In this case we will
be accounting the memcg allocation to the root cgroup until we succeed to
create its own cache, but it isn't that critical.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
kmem_cache_dup() is only called from memcg_create_kmem_cache(). The
latter, in fact, does nothing besides this, so let's fold kmem_cache_dup()
into memcg_create_kmem_cache().
This patch also makes the memcg_cache_mutex private to
memcg_create_kmem_cache(), because it is not used anywhere else.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
We obtain a per-memcg cache from a root kmem_cache by dereferencing an
entry of the root cache's memcg_params::memcg_caches array. If we find no
cache for a memcg there on allocation, we initiate the memcg cache
creation (see memcg_kmem_get_cache()). The cache creation proceeds
asynchronously in memcg_create_kmem_cache() in order to avoid lock
clashes, so there can be several threads trying to create the same
kmem_cache concurrently, but only one of them may succeed. However, due
to a race in the code, it is not always true. The point is that the
memcg_caches array can be relocated when we activate kmem accounting for a
memcg (see memcg_update_all_caches(), memcg_update_cache_size()). If
memcg_update_cache_size() and memcg_create_kmem_cache() proceed
concurrently as described below, we can leak a kmem_cache.
Asume two threads schedule creation of the same kmem_cache. One of them
successfully creates it. Another one should fail then, but if
memcg_create_kmem_cache() interleaves with memcg_update_cache_size() as
follows, it won't:
memcg_create_kmem_cache() memcg_update_cache_size()
(called w/o mutexes held) (called with slab_mutex,
set_limit_mutex held)
------------------------- -------------------------
mutex_lock(&memcg_cache_mutex)
s->memcg_params=kzalloc(...)
new_cachep=cache_from_memcg_idx(cachep,idx)
// new_cachep==NULL => proceed to creation
s->memcg_params->memcg_caches[i]
=cur_params->memcg_caches[i]
// kmem_cache_create_memcg takes slab_mutex
// so we will hang around until
// memcg_update_cache_size finishes, but
// nothing will prevent it from succeeding so
// memcg_caches[idx] will be overwritten in
// memcg_register_cache!
new_cachep = kmem_cache_create_memcg(...)
mutex_unlock(&memcg_cache_mutex)
Let's fix this by moving the check for existence of the memcg cache to
kmem_cache_create_memcg() to be called under the slab_mutex and make it
return NULL if so.
A similar race is possible when destroying a memcg cache (see
kmem_cache_destroy()). Since memcg_unregister_cache(), which clears the
pointer in the memcg_caches array, is called w/o protection, we can race
with memcg_update_cache_size() and omit clearing the pointer. Therefore
memcg_unregister_cache() should be moved before we release the slab_mutex.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
All caches of the same memory cgroup are linked in the memcg_slab_caches
list via kmem_cache::memcg_params::list. This list is traversed, for
example, when we read memory.kmem.slabinfo. Since the list actually
consists of memcg_cache_params objects, we have to convert an element of
the list to a kmem_cache object using memcg_params_to_cache(), which
obtains the pointer to the cache from the memcg_params::memcg_caches array
of the corresponding root cache. That said the pointer to a kmem_cache in
its parent's memcg_params must be initialized before adding the cache to
the list, and cleared only after it has been unlinked. Currently it is
vice-versa, which can result in a NULL ptr dereference while traversing
the memcg_slab_caches list. This patch restores the correct order.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Each root kmem_cache has pointers to per-memcg caches stored in its
memcg_params::memcg_caches array. Whenever we want to allocate a slab for
a memcg, we access this array to get per-memcg cache to allocate from (see
memcg_kmem_get_cache()). The access must be lock-free for performance
reasons, so we should use barriers to assert the kmem_cache is up-to-date.
First, we should place a write barrier immediately before setting the
pointer to it in the memcg_caches array in order to make sure nobody will
see a partially initialized object. Second, we should issue a read
barrier before dereferencing the pointer to conform to the write barrier.
However, currently the barrier usage looks rather strange. We have a
write barrier *after* setting the pointer and a read barrier *before*
reading the pointer, which is incorrect. This patch fixes this.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently, we have rather a messy function set relating to per-memcg kmem
cache initialization/destruction.
Per-memcg caches are created in memcg_create_kmem_cache(). This function
calls kmem_cache_create_memcg() to allocate and initialize a kmem cache
and then "registers" the new cache in the memcg_params::memcg_caches array
of the parent cache.
During its work-flow, kmem_cache_create_memcg() executes the following
memcg-related functions:
- memcg_alloc_cache_params(), to initialize memcg_params of the newly
created cache;
- memcg_cache_list_add(), to add the new cache to the memcg_slab_caches
list.
On the other hand, kmem_cache_destroy() called on a cache destruction only
calls memcg_release_cache(), which does all the work: it cleans the
reference to the cache in its parent's memcg_params::memcg_caches, removes
the cache from the memcg_slab_caches list, and frees memcg_params.
Such an inconsistency between destruction and initialization paths make
the code difficult to read, so let's clean this up a bit.
This patch moves all the code relating to registration of per-memcg caches
(adding to memcg list, setting the pointer to a cache from its parent) to
the newly created memcg_register_cache() and memcg_unregister_cache()
functions making the initialization and destruction paths look
symmetrical.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
We do not free the cache's memcg_params if __kmem_cache_create fails. Fix
this.
Plus, rename memcg_register_cache() to memcg_alloc_cache_params(), because
it actually does not register the cache anywhere, but simply initialize
kmem_cache::memcg_params.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently kmem_cache_create_memcg() backoffs on failure inside
conditionals, without using gotos. This results in the rollback code
duplication, which makes the function look cumbersome even though on error
we should only free the allocated cache. Since in the next patch I am
going to add yet another rollback function call on error path there, let's
employ labels instead of conditionals for undoing any changes on failure
to keep things clean.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Reviewed-by: Pekka Enberg <penberg@kernel.org>
Cc: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Christoph Lameter <cl@linux.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
get/put_page(thp_tail) paths do get_page_unless_zero(page_head) +
compound_lock(). In theory this page_head can be already freed and
reallocated as alloc_pages(__GFP_COMP, smaller_order). In this case
get_page_unless_zero() can succeed right after set_page_refcounted(), and
compound_lock() can race with the non-atomic __SetPageHead() in
prep_compound_page().
Perhaps we should rework the thp locking (under discussion), but until
then this patch moves set_page_refcounted() and adds wmb() to ensure that
page->_count != 0 comes as a last change.
I am not sure about other callers of set_page_refcounted(), but at first
glance they look fine to me.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Most of the VM_BUG_ON assertions are performed on a page. Usually, when
one of these assertions fails we'll get a BUG_ON with a call stack and the
registers.
I've recently noticed based on the requests to add a small piece of code
that dumps the page to various VM_BUG_ON sites that the page dump is quite
useful to people debugging issues in mm.
This patch adds a VM_BUG_ON_PAGE(cond, page) which beyond doing what
VM_BUG_ON() does, also dumps the page before executing the actual BUG_ON.
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The vmalloc was introduced by 333279 ("memcgroup: use vmalloc for
mem_cgroup allocation"), because at that time MAX_NUMNODES was used for
defining the per-node array in the mem_cgroup structure so that the
structure could be huge even if the system had the only NUMA node.
The situation was significantly improved by patch 45cf7e ("memcg: reduce
the size of struct memcg 244-fold"), which made the size of the mem_cgroup
structure calculated dynamically depending on the real number of NUMA
nodes installed on the system (nr_node_ids), so now there is no point in
using vmalloc here: the structure is allocated rarely and on most systems
its size is about 1K.
Signed-off-by: Vladimir Davydov <vdavydov@parallels.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Cc: Glauber Costa <glommer@openvz.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
avoid a coding-style ugly
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Since commit ff6a6da60 ("mm: accelerate munlock() treatment of THP pages")
munlock skips tail pages of a munlocked THP page. There is some attempt
to prevent bad consequences of racing with a THP page split, but code
inspection indicates that there are two problems that may lead to a
non-fatal, yet wrong outcome.
First, __split_huge_page_refcount() copies flags including PageMlocked
from the head page to the tail pages. Clearing PageMlocked by
munlock_vma_page() in the middle of this operation might result in part of
tail pages left with PageMlocked flag. As the head page still appears to
be a THP page until all tail pages are processed, munlock_vma_page() might
think it munlocked the whole THP page and skip all the former tail pages.
Before ff6a6da60, those pages would be cleared in further iterations of
munlock_vma_pages_range(), but NR_MLOCK would still become undercounted
(related the next point).
Second, NR_MLOCK accounting is based on call to hpage_nr_pages() after the
PageMlocked is cleared. The accounting might also become inconsistent due
to race with __split_huge_page_refcount()
- undercount when HUGE_PMD_NR is subtracted, but some tail pages are
left with PageMlocked set and counted again (only possible before
ff6a6da60)
- overcount when hpage_nr_pages() sees a normal page (split has already
finished), but the parallel split has meanwhile cleared PageMlocked from
additional tail pages
This patch prevents both problems via extending the scope of lru_lock in
munlock_vma_page(). This is convenient because:
- __split_huge_page_refcount() takes lru_lock for its whole operation
- munlock_vma_page() typically takes lru_lock anyway for page isolation
As this becomes a second function where page isolation is done with
lru_lock already held, factor this out to a new
__munlock_isolate_lru_page() function and clean up the code around.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Hugh Dickins <hughd@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
switch to pr_alert.
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
bad_page() is cool in that it prints out a bunch of data about the page.
But, I can never remember which page flags are good and which are bad, or
whether ->index or ->mapping is required to be NULL.
This patch allows bad/dump_page() callers to specify a string about why
they are dumping the page and adds explanation strings to a number of
places. It also adds a 'bad_flags' argument to bad_page(), which it then
dumps out separately from the flags which are actually set.
This way, the messages will show specifically why the page was bad,
*specifically* which flags it is complaining about, if it was a page flag
combination which was the problem.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Christoph Lameter <cl@linux.com>
Cc: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The "compressor" and "enabled" params are currently hidden, this changes
them to read-only, so userspace can tell if zswap is enabled or not and
see what compressor is in use.
Signed-off-by: Dan Streetman <ddstreet@ieee.org>
Cc: Vladimir Murzin <murzin.v@gmail.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Weijie Yang <weijie.yang@samsung.com>
Acked-by: Seth Jennings <sjennings@variantweb.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
fail_migrate_page() isn't used anywhere, so remove it.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Rafael Aquini <aquini@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Some part of putback_lru_pages() and putback_movable_pages() is
duplicated, so it could confuse us what we should use. We can remove
putback_lru_pages() since it is not really needed now. This makes us
undestand and maintain the code more easily.
And comment on putback_movable_pages() is stale now, so fix it.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Rafael Aquini <aquini@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
We should remove the page from the list if we fail with ENOSYS, since
migrate_pages() consider error cases except -ENOMEM and -EAGAIN as
permanent failure and it assumes that the page would be removed from the
list. Without this patch, we could overcount number of failure.
In addition, we should put back the new hugepage if
!hugepage_migration_support(). If not, we would leak hugepage memory.
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Rafael Aquini <aquini@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Let's add a comment about where the failed page goes to, which makes code
more readable.
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Acked-by: Christoph Lameter <cl@linux.com>
Reviewed-by: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Acked-by: Rafael Aquini <aquini@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wanpeng Li <liwanp@linux.vnet.ibm.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zhang Yanfei <zhangyanfei@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
__GFP_NOFAIL may return NULL when coupled with GFP_NOWAIT or GFP_ATOMIC.
Luckily, nothing currently does such craziness. So instead of causing
such allocations to loop (potentially forever), we maintain the current
behavior and also warn about the new users of the deprecated flag.
Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Compaction used to start its migrate and free page scaners at the zone's
lowest and highest pfn, respectively. Later, caching was introduced to
remember the scanners' progress across compaction attempts so that
pageblocks are not re-scanned uselessly. Additionally, pageblocks where
isolation failed are marked to be quickly skipped when encountered again
in future compactions.
Currently, both the reset of cached pfn's and clearing of the pageblock
skip information for a zone is done in __reset_isolation_suitable(). This
function gets called when:
- compaction is restarting after being deferred
- compact_blockskip_flush flag is set in compact_finished() when the scanners
meet (and not again cleared when direct compaction succeeds in allocation)
and kswapd acts upon this flag before going to sleep
This behavior is suboptimal for several reasons:
- when direct sync compaction is called after async compaction fails (in the
allocation slowpath), it will effectively do nothing, unless kswapd
happens to process the compact_blockskip_flush flag meanwhile. This is racy
and goes against the purpose of sync compaction to more thoroughly retry
the compaction of a zone where async compaction has failed.
The restart-after-deferring path cannot help here as deferring happens only
after the sync compaction fails. It is also done only for the preferred
zone, while the compaction might be done for a fallback zone.
- the mechanism of marking pageblock to be skipped has little value since the
cached pfn's are reset only together with the pageblock skip flags. This
effectively limits pageblock skip usage to parallel compactions.
This patch changes compact_finished() so that cached pfn's are reset
immediately when the scanners meet. Clearing pageblock skip flags is
unchanged, as well as the other situations where cached pfn's are reset.
This allows the sync-after-async compaction to retry pageblocks not marked
as skipped, such as blocks !MIGRATE_MOVABLE blocks that async compactions
now skips without marking them.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Rik van Riel <riel@redhat.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Compaction temporarily marks pageblocks where it fails to isolate pages as
to-be-skipped in further compactions, in order to improve efficiency. One
of the reasons to fail isolating pages is that isolation is not attempted
in pageblocks that are not of MIGRATE_MOVABLE (or CMA) type.
The problem is that blocks skipped due to not being MIGRATE_MOVABLE in
async compaction become skipped due to the temporary mark also in future
sync compaction. Moreover, this may follow quite soon during
__alloc_page_slowpath, without much time for kswapd to clear the pageblock
skip marks. This goes against the idea that sync compaction should try to
scan these blocks more thoroughly than the async compaction.
The fix is to ensure in async compaction that these !MIGRATE_MOVABLE
blocks are not marked to be skipped. Note this should not affect
performance or locking impact of further async compactions, as skipping a
block due to being !MIGRATE_MOVABLE is done soon after skipping a block
marked to be skipped, both without locking.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Rik van Riel <riel@redhat.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Compaction of a zone is finished when the migrate scanner (which begins at
the zone's lowest pfn) meets the free page scanner (which begins at the
zone's highest pfn). This is detected in compact_zone() and in the case
of direct compaction, the compact_blockskip_flush flag is set so that
kswapd later resets the cached scanner pfn's, and a new compaction may
again start at the zone's borders.
The meeting of the scanners can happen during either scanner's activity.
However, it may currently fail to be detected when it occurs in the free
page scanner, due to two problems. First, isolate_freepages() keeps
free_pfn at the highest block where it isolated pages from, for the
purposes of not missing the pages that are returned back to allocator when
migration fails. Second, failing to isolate enough free pages due to
scanners meeting results in -ENOMEM being returned by migrate_pages(),
which makes compact_zone() bail out immediately without calling
compact_finished() that would detect scanners meeting.
This failure to detect scanners meeting might result in repeated attempts
at compaction of a zone that keep starting from the cached pfn's close to
the meeting point, and quickly failing through the -ENOMEM path, without
the cached pfns being reset, over and over. This has been observed
(through additional tracepoints) in the third phase of the mmtests
stress-highalloc benchmark, where the allocator runs on an otherwise idle
system. The problem was observed in the DMA32 zone, which was used as a
fallback to the preferred Normal zone, but on the 4GB system it was
actually the largest zone. The problem is even amplified for such
fallback zone - the deferred compaction logic, which could (after being
fixed by a previous patch) reset the cached scanner pfn's, is only applied
to the preferred zone and not for the fallbacks.
The problem in the third phase of the benchmark was further amplified by
commit 81c0a2bb ("mm: page_alloc: fair zone allocator policy") which
resulted in a non-deterministic regression of the allocation success rate
from ~85% to ~65%. This occurs in about half of benchmark runs, making
bisection problematic. It is unlikely that the commit itself is buggy,
but it should put more pressure on the DMA32 zone during phases 1 and 2,
which may leave it more fragmented in phase 3 and expose the bugs that
this patch fixes.
The fix is to make scanners meeting in isolate_freepage() stay that way,
and to check in compact_zone() for scanners meeting when migrate_pages()
returns -ENOMEM. The result is that compact_finished() also detects
scanners meeting and sets the compact_blockskip_flush flag to make kswapd
reset the scanner pfn's.
The results in stress-highalloc benchmark show that the "regression" by
commit 81c0a2bb in phase 3 no longer occurs, and phase 1 and 2 allocation
success rates are also significantly improved.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Rik van Riel <riel@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Compaction caches pfn's for its migrate and free scanners to avoid
scanning the whole zone each time. In compact_zone(), the cached values
are read to set up initial values for the scanners. There are several
situations when these cached pfn's are reset to the first and last pfn of
the zone, respectively. One of these situations is when a compaction has
been deferred for a zone and is now being restarted during a direct
compaction, which is also done in compact_zone().
However, compact_zone() currently reads the cached pfn's *before*
resetting them. This means the reset doesn't affect the compaction that
performs it, and with good chance also subsequent compactions, as
update_pageblock_skip() is likely to be called and update the cached pfn's
to those being processed. Another chance for a successful reset is when a
direct compaction detects that migration and free scanners meet (which has
its own problems addressed by another patch) and sets
update_pageblock_skip flag which kswapd uses to do the reset because it
goes to sleep.
This is clearly a bug that results in non-deterministic behavior, so this
patch moves the cached pfn reset to be performed *before* the values are
read.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Mel Gorman <mgorman@suse.de>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently there are several functions to manipulate the deferred
compaction state variables. The remaining case where the variables are
touched directly is when a successful allocation occurs in direct
compaction, or is expected to be successful in the future by kswapd.
Here, the lowest order that is expected to fail is updated, and in the
case of successful allocation, the deferred status and counter is reset
completely.
Create a new function compaction_defer_reset() to encapsulate this
functionality and make it easier to understand the code. No functional
change.
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The broad goal of the series is to improve allocation success rates for
huge pages through memory compaction, while trying not to increase the
compaction overhead. The original objective was to reintroduce capturing
of high-order pages freed by the compaction, before they are split by
concurrent activity. However, several bugs and opportunities for simple
improvements were found in the current implementation, mostly through
extra tracepoints (which are however too ugly for now to be considered for
sending).
The patches mostly deal with two mechanisms that reduce compaction
overhead, which is caching the progress of migrate and free scanners, and
marking pageblocks where isolation failed to be skipped during further
scans.
Patch 1 (from mgorman) adds tracepoints that allow calculate time spent in
compaction and potentially debug scanner pfn values.
Patch 2 encapsulates the some functionality for handling deferred compactions
for better maintainability, without a functional change
type is not determined without being actually needed.
Patch 3 fixes a bug where cached scanner pfn's are sometimes reset only after
they have been read to initialize a compaction run.
Patch 4 fixes a bug where scanners meeting is sometimes not properly detected
and can lead to multiple compaction attempts quitting early without
doing any work.
Patch 5 improves the chances of sync compaction to process pageblocks that
async compaction has skipped due to being !MIGRATE_MOVABLE.
Patch 6 improves the chances of sync direct compaction to actually do anything
when called after async compaction fails during allocation slowpath.
The impact of patches were validated using mmtests's stress-highalloc
benchmark with mmtests's stress-highalloc benchmark on a x86_64 machine
with 4GB memory.
Due to instability of the results (mostly related to the bugs fixed by
patches 2 and 3), 10 iterations were performed, taking min,mean,max values
for success rates and mean values for time and vmstat-based metrics.
First, the default GFP_HIGHUSER_MOVABLE allocations were tested with the
patches stacked on top of v3.13-rc2. Patch 2 is OK to serve as baseline
due to no functional changes in 1 and 2. Comments below.
stress-highalloc
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-nothp 3-nothp 4-nothp 5-nothp 6-nothp
Success 1 Min 9.00 ( 0.00%) 10.00 (-11.11%) 43.00 (-377.78%) 43.00 (-377.78%) 33.00 (-266.67%)
Success 1 Mean 27.50 ( 0.00%) 25.30 ( 8.00%) 45.50 (-65.45%) 45.90 (-66.91%) 46.30 (-68.36%)
Success 1 Max 36.00 ( 0.00%) 36.00 ( 0.00%) 47.00 (-30.56%) 48.00 (-33.33%) 52.00 (-44.44%)
Success 2 Min 10.00 ( 0.00%) 8.00 ( 20.00%) 46.00 (-360.00%) 45.00 (-350.00%) 35.00 (-250.00%)
Success 2 Mean 26.40 ( 0.00%) 23.50 ( 10.98%) 47.30 (-79.17%) 47.60 (-80.30%) 48.10 (-82.20%)
Success 2 Max 34.00 ( 0.00%) 33.00 ( 2.94%) 48.00 (-41.18%) 50.00 (-47.06%) 54.00 (-58.82%)
Success 3 Min 65.00 ( 0.00%) 63.00 ( 3.08%) 85.00 (-30.77%) 84.00 (-29.23%) 85.00 (-30.77%)
Success 3 Mean 76.70 ( 0.00%) 70.50 ( 8.08%) 86.20 (-12.39%) 85.50 (-11.47%) 86.00 (-12.13%)
Success 3 Max 87.00 ( 0.00%) 86.00 ( 1.15%) 88.00 ( -1.15%) 87.00 ( 0.00%) 87.00 ( 0.00%)
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-nothp 3-nothp 4-nothp 5-nothp 6-nothp
User 6437.72 6459.76 5960.32 5974.55 6019.67
System 1049.65 1049.09 1029.32 1031.47 1032.31
Elapsed 1856.77 1874.48 1949.97 1994.22 1983.15
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-nothp 3-nothp 4-nothp 5-nothp 6-nothp
Minor Faults 253952267 254581900 250030122 250507333 250157829
Major Faults 420 407 506 530 530
Swap Ins 4 9 9 6 6
Swap Outs 398 375 345 346 333
Direct pages scanned 197538 189017 298574 287019 299063
Kswapd pages scanned 1809843 1801308 1846674 1873184 1861089
Kswapd pages reclaimed 1806972 1798684 1844219 1870509 1858622
Direct pages reclaimed 197227 188829 298380 286822 298835
Kswapd efficiency 99% 99% 99% 99% 99%
Kswapd velocity 953.382 970.449 952.243 934.569 922.286
Direct efficiency 99% 99% 99% 99% 99%
Direct velocity 104.058 101.832 153.961 143.200 148.205
Percentage direct scans 9% 9% 13% 13% 13%
Zone normal velocity 347.289 359.676 348.063 339.933 332.983
Zone dma32 velocity 710.151 712.605 758.140 737.835 737.507
Zone dma velocity 0.000 0.000 0.000 0.000 0.000
Page writes by reclaim 557.600 429.000 353.600 426.400 381.800
Page writes file 159 53 7 79 48
Page writes anon 398 375 345 346 333
Page reclaim immediate 825 644 411 575 420
Sector Reads 2781750 2769780 2878547 2939128 2910483
Sector Writes 12080843 12083351 12012892 12002132 12010745
Page rescued immediate 0 0 0 0 0
Slabs scanned 1575654 1545344 1778406 1786700 1794073
Direct inode steals 9657 10037 15795 14104 14645
Kswapd inode steals 46857 46335 50543 50716 51796
Kswapd skipped wait 0 0 0 0 0
THP fault alloc 97 91 81 71 77
THP collapse alloc 456 506 546 544 565
THP splits 6 5 5 4 4
THP fault fallback 0 1 0 0 0
THP collapse fail 14 14 12 13 12
Compaction stalls 1006 980 1537 1536 1548
Compaction success 303 284 562 559 578
Compaction failures 702 696 974 976 969
Page migrate success 1177325 1070077 3927538 3781870 3877057
Page migrate failure 0 0 0 0 0
Compaction pages isolated 2547248 2306457 8301218 8008500 8200674
Compaction migrate scanned 42290478 38832618 153961130 154143900 159141197
Compaction free scanned 89199429 79189151 356529027 351943166 356326727
Compaction cost 1566 1426 5312 5156 5294
NUMA PTE updates 0 0 0 0 0
NUMA hint faults 0 0 0 0 0
NUMA hint local faults 0 0 0 0 0
NUMA hint local percent 100 100 100 100 100
NUMA pages migrated 0 0 0 0 0
AutoNUMA cost 0 0 0 0 0
Observations:
- The "Success 3" line is allocation success rate with system idle
(phases 1 and 2 are with background interference). I used to get stable
values around 85% with vanilla 3.11. The lower min and mean values came
with 3.12. This was bisected to commit 81c0a2bb ("mm: page_alloc: fair
zone allocator policy") As explained in comment for patch 3, I don't
think the commit is wrong, but that it makes the effect of compaction
bugs worse. From patch 3 onwards, the results are OK and match the 3.11
results.
- Patch 4 also clearly helps phases 1 and 2, and exceeds any results
I've seen with 3.11 (I didn't measure it that thoroughly then, but it
was never above 40%).
- Compaction cost and number of scanned pages is higher, especially due
to patch 4. However, keep in mind that patches 3 and 4 fix existing
bugs in the current design of compaction overhead mitigation, they do
not change it. If overhead is found unacceptable, then it should be
decreased differently (and consistently, not due to random conditions)
than the current implementation does. In contrast, patches 5 and 6
(which are not strictly bug fixes) do not increase the overhead (but
also not success rates). This might be a limitation of the
stress-highalloc benchmark as it's quite uniform.
Another set of results is when configuring stress-highalloc t allocate
with similar flags as THP uses:
(GFP_HIGHUSER_MOVABLE|__GFP_NOMEMALLOC|__GFP_NORETRY|__GFP_NO_KSWAPD)
stress-highalloc
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-thp 3-thp 4-thp 5-thp 6-thp
Success 1 Min 2.00 ( 0.00%) 7.00 (-250.00%) 18.00 (-800.00%) 19.00 (-850.00%) 26.00 (-1200.00%)
Success 1 Mean 19.20 ( 0.00%) 17.80 ( 7.29%) 29.20 (-52.08%) 29.90 (-55.73%) 32.80 (-70.83%)
Success 1 Max 27.00 ( 0.00%) 29.00 ( -7.41%) 35.00 (-29.63%) 36.00 (-33.33%) 37.00 (-37.04%)
Success 2 Min 3.00 ( 0.00%) 8.00 (-166.67%) 21.00 (-600.00%) 21.00 (-600.00%) 32.00 (-966.67%)
Success 2 Mean 19.30 ( 0.00%) 17.90 ( 7.25%) 32.20 (-66.84%) 32.60 (-68.91%) 35.70 (-84.97%)
Success 2 Max 27.00 ( 0.00%) 30.00 (-11.11%) 36.00 (-33.33%) 37.00 (-37.04%) 39.00 (-44.44%)
Success 3 Min 62.00 ( 0.00%) 62.00 ( 0.00%) 85.00 (-37.10%) 75.00 (-20.97%) 64.00 ( -3.23%)
Success 3 Mean 66.30 ( 0.00%) 65.50 ( 1.21%) 85.60 (-29.11%) 83.40 (-25.79%) 83.50 (-25.94%)
Success 3 Max 70.00 ( 0.00%) 69.00 ( 1.43%) 87.00 (-24.29%) 86.00 (-22.86%) 87.00 (-24.29%)
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-thp 3-thp 4-thp 5-thp 6-thp
User 6547.93 6475.85 6265.54 6289.46 6189.96
System 1053.42 1047.28 1043.23 1042.73 1038.73
Elapsed 1835.43 1821.96 1908.67 1912.74 1956.38
3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2 3.13-rc2
2-thp 3-thp 4-thp 5-thp 6-thp
Minor Faults 256805673 253106328 253222299 249830289 251184418
Major Faults 395 375 423 434 448
Swap Ins 12 10 10 12 9
Swap Outs 530 537 487 455 415
Direct pages scanned 71859 86046 153244 152764 190713
Kswapd pages scanned 1900994 1870240 1898012 1892864 1880520
Kswapd pages reclaimed 1897814 1867428 1894939 1890125 1877924
Direct pages reclaimed 71766 85908 153167 152643 190600
Kswapd efficiency 99% 99% 99% 99% 99%
Kswapd velocity 1029.000 1067.782 1000.091 991.049 951.218
Direct efficiency 99% 99% 99% 99% 99%
Direct velocity 38.897 49.127 80.747 79.983 96.468
Percentage direct scans 3% 4% 7% 7% 9%
Zone normal velocity 351.377 372.494 348.910 341.689 335.310
Zone dma32 velocity 716.520 744.414 731.928 729.343 712.377
Zone dma velocity 0.000 0.000 0.000 0.000 0.000
Page writes by reclaim 669.300 604.000 545.700 538.900 429.900
Page writes file 138 66 58 83 14
Page writes anon 530 537 487 455 415
Page reclaim immediate 806 655 772 548 517
Sector Reads 2711956 2703239 2811602 2818248 2839459
Sector Writes 12163238 12018662 12038248 11954736 11994892
Page rescued immediate 0 0 0 0 0
Slabs scanned 1385088 1388364 1507968 1513292 1558656
Direct inode steals 1739 2564 4622 5496 6007
Kswapd inode steals 47461 46406 47804 48013 48466
Kswapd skipped wait 0 0 0 0 0
THP fault alloc 110 82 84 69 70
THP collapse alloc 445 482 467 462 539
THP splits 6 5 4 5 3
THP fault fallback 3 0 0 0 0
THP collapse fail 15 14 14 14 13
Compaction stalls 659 685 1033 1073 1111
Compaction success 222 225 410 427 456
Compaction failures 436 460 622 646 655
Page migrate success 446594 439978 1085640 1095062 1131716
Page migrate failure 0 0 0 0 0
Compaction pages isolated 1029475 1013490 2453074 2482698 2565400
Compaction migrate scanned 9955461 11344259 24375202 27978356 30494204
Compaction free scanned 27715272 28544654 80150615 82898631 85756132
Compaction cost 552 555 1344 1379 1436
NUMA PTE updates 0 0 0 0 0
NUMA hint faults 0 0 0 0 0
NUMA hint local faults 0 0 0 0 0
NUMA hint local percent 100 100 100 100 100
NUMA pages migrated 0 0 0 0 0
AutoNUMA cost 0 0 0 0 0
There are some differences from the previous results for THP-like allocations:
- Here, the bad result for unpatched kernel in phase 3 is much more
consistent to be between 65-70% and not related to the "regression" in
3.12. Still there is the improvement from patch 4 onwards, which brings
it on par with simple GFP_HIGHUSER_MOVABLE allocations.
- Compaction costs have increased, but nowhere near as much as the
non-THP case. Again, the patches should be worth the gained
determininsm.
- Patches 5 and 6 somewhat increase the number of migrate-scanned pages.
This is most likely due to __GFP_NO_KSWAPD flag, which means the cached
pfn's and pageblock skip bits are not reset by kswapd that often (at
least in phase 3 where no concurrent activity would wake up kswapd) and
the patches thus help the sync-after-async compaction. It doesn't
however show that the sync compaction would help so much with success
rates, which can be again seen as a limitation of the benchmark
scenario.
This patch (of 6):
Add two tracepoints for compaction begin and end of a zone. Using this it
is possible to calculate how much time a workload is spending within
compaction and potentially debug problems related to cached pfns for
scanning. In combination with the direct reclaim and slab trace points it
should be possible to estimate most allocation-related overhead for a
workload.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Rik van Riel <riel@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
