Patch series "Enable THP for text section of non-shmem files", v10;
This patchset follows up discussion at LSF/MM 2019. The motivation is to
put text section of an application in THP, and thus reduces iTLB miss rate
and improves performance. Both Facebook and Oracle showed strong
interests to this feature.
To make reviews easier, this set aims a mininal valid product. Current
version of the work does not have any changes to file system specific
code. This comes with some limitations (discussed later).
This set enables an application to "hugify" its text section by simply
running something like:
madvise(0x600000, 0x80000, MADV_HUGEPAGE);
Before this call, the /proc/<pid>/maps looks like:
00400000-074d0000 r-xp 00000000 00:27 2006927 app
After this call, part of the text section is split out and mapped to
THP:
00400000-00425000 r-xp 00000000 00:27 2006927 app
00600000-00e00000 r-xp 00200000 00:27 2006927 app <<< on THP
00e00000-074d0000 r-xp 00a00000 00:27 2006927 app
Limitations:
1. This only works for text section (vma with VM_DENYWRITE).
2. Original limitation #2 is removed in v3.
We gated this feature with an experimental config, READ_ONLY_THP_FOR_FS.
Once we get better support on the write path, we can remove the config and
enable it by default.
Tested cases:
1. Tested with btrfs and ext4.
2. Tested with real work application (memcache like caching service).
3. Tested with "THP aware uprobe":
https://patchwork.kernel.org/project/linux-mm/list/?series=131339
This patch (of 7):
Currently, filemap_fault() avoids race condition with truncate by checking
page->mapping == mapping. This does not work for compound pages. This
patch let it check compound_head(page)->mapping instead.
Link: http://lkml.kernel.org/r/20190801184244.3169074-2-songliubraving@fb.com
Signed-off-by: Song Liu <songliubraving@fb.com>
Acked-by: Rik van Riel <riel@surriel.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When allocating hugetlbfs pool pages via /proc/sys/vm/nr_hugepages, the
pages will be interleaved between all nodes of the system. If nodes are
not equal, it is quite possible for one node to fill up before the others.
When this happens, the code still attempts to allocate pages from the
full node. This results in calls to direct reclaim and compaction which
slow things down considerably.
When allocating pool pages, note the state of the previous allocation for
each node. If previous allocation failed, do not use the aggressive retry
algorithm on successive attempts. The allocation will still succeed if
there is memory available, but it will not try as hard to free up memory.
Link: http://lkml.kernel.org/r/20190806014744.15446-5-mike.kravetz@oracle.com
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mike Kravetz reports that "hugetlb allocations could stall for minutes or
hours when should_compact_retry() would return true more often then it
should. Specifically, this was in the case where compact_result was
COMPACT_DEFERRED and COMPACT_PARTIAL_SKIPPED and no progress was being
made."
The problem is that the compaction_withdrawn() test in
should_compact_retry() includes compaction outcomes that are only possible
on low compaction priority, and results in a retry without increasing the
priority. This may result in furter reclaim, and more incomplete
compaction attempts.
With this patch, compaction priority is raised when possible, or
should_compact_retry() returns false.
The COMPACT_SKIPPED result doesn't really fit together with the other
outcomes in compaction_withdrawn(), as that's a result caused by
insufficient order-0 pages, not due to low compaction priority. With this
patch, it is moved to a new compaction_needs_reclaim() function, and for
that outcome we keep the current logic of retrying if it looks like
reclaim will be able to help.
Link: http://lkml.kernel.org/r/20190806014744.15446-4-mike.kravetz@oracle.com
Reported-by: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Tested-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After commit "mm, reclaim: make should_continue_reclaim perform dryrun
detection", closer look at the function shows, that nr_reclaimed == 0
means the function will always return false. And since non-zero
nr_reclaimed implies non_zero nr_scanned, testing nr_scanned serves no
purpose, and so does the testing for __GFP_RETRY_MAYFAIL.
This patch thus cleans up the function to test only !nr_reclaimed upfront,
and remove the __GFP_RETRY_MAYFAIL test and nr_scanned parameter
completely. Comment is also updated, explaining that approximating "full
LRU list has been scanned" with nr_scanned == 0 didn't really work.
Link: http://lkml.kernel.org/r/20190806014744.15446-3-mike.kravetz@oracle.com
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "address hugetlb page allocation stalls", v2.
Allocation of hugetlb pages via sysctl or procfs can stall for minutes or
hours. A simple example on a two node system with 8GB of memory is as
follows:
echo 4096 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages
echo 4096 > /proc/sys/vm/nr_hugepages
Obviously, both allocation attempts will fall short of their 8GB goal.
However, one or both of these commands may stall and not be interruptible.
The issues were initially discussed in mail thread [1] and RFC code at
[2].
This series addresses the issues causing the stalls. There are two
distinct fixes, a cleanup, and an optimization. The reclaim patch by
Hillf and compaction patch by Vlasitmil address corner cases in their
respective areas. hugetlb page allocation could stall due to either of
these issues. Vlasitmil added a cleanup patch after Hillf's
modifications. The hugetlb patch by Mike is an optimization suggested
during the debug and development process.
[1] http://lkml.kernel.org/r/d38a095e-dc39-7e82-bb76-2c9247929f07@oracle.com
[2] http://lkml.kernel.org/r/20190724175014.9935-1-mike.kravetz@oracle.com
This patch (of 4):
Address the issue of should_continue_reclaim returning true too often for
__GFP_RETRY_MAYFAIL attempts when !nr_reclaimed and nr_scanned. This was
observed during hugetlb page allocation causing stalls for minutes or
hours.
We can stop reclaiming pages if compaction reports it can make a progress.
There might be side-effects for other high-order allocations that would
potentially benefit from reclaiming more before compaction so that they
would be faster and less likely to stall. However, the consequences of
premature/over-reclaim are considered worse.
We can also bail out of reclaiming pages if we know that there are not
enough inactive lru pages left to satisfy the costly allocation.
We can give up reclaiming pages too if we see dryrun occur, with the
certainty of plenty of inactive pages. IOW with dryrun detected, we are
sure we have reclaimed as many pages as we could.
Link: http://lkml.kernel.org/r/20190806014744.15446-2-mike.kravetz@oracle.com
Signed-off-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Tested-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Cgroup v1 memcg controller has exposed a dedicated kmem limit to users
which turned out to be really a bad idea because there are paths which
cannot shrink the kernel memory usage enough to get below the limit (e.g.
because the accounted memory is not reclaimable). There are cases when
the failure is even not allowed (e.g. __GFP_NOFAIL). This means that the
kmem limit is in excess to the hard limit without any way to shrink and
thus completely useless. OOM killer cannot be invoked to handle the
situation because that would lead to a premature oom killing.
As a result many places might see ENOMEM returning from kmalloc and result
in unexpected errors. E.g. a global OOM killer when there is a lot of
free memory because ENOMEM is translated into VM_FAULT_OOM in #PF path and
therefore pagefault_out_of_memory would result in OOM killer.
Please note that the kernel memory is still accounted to the overall limit
along with the user memory so removing the kmem specific limit should
still allow to contain kernel memory consumption. Unlike the kmem one,
though, it invokes memory reclaim and targeted memcg oom killing if
necessary.
Start the deprecation process by crying to the kernel log. Let's see
whether there are relevant usecases and simply return to EINVAL in the
second stage if nobody complains in few releases.
[akpm@linux-foundation.org: tweak documentation text]
Link: http://lkml.kernel.org/r/20190911151612.GI4023@dhcp22.suse.cz
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Thomas Lindroth <thomas.lindroth@gmail.com>
Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mem_cgroup_id_get() was introduced in commit 73f576c04b ("mm:memcontrol:
fix cgroup creation failure after many small jobs").
Later, it no longer has any user since the commits,
1f47b61fb4 ("mm: memcontrol: fix swap counter leak on swapout from offline cgroup")
58fa2a5512 ("mm: memcontrol: add sanity checks for memcg->id.ref on get/put")
so safe to remove it.
Link: http://lkml.kernel.org/r/1568648453-5482-1-git-send-email-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
constrained_alloc() calculates the size of the oom domain by using
node_spanned_pages which is incorrect because this is the full range of
the physical memory range that the numa node occupies rather than the
memory that backs that range which is represented by node_present_pages.
Sparsely populated nodes (e.g. after memory hot remove or simply sparse
due to memory layout) can have really a large difference between the two.
This shouldn't really cause any real user observable problems because the
oom calculates a ratio against totalpages and used memory cannot exceed
present pages but it is confusing and wrong from code point of view.
Link: http://lkml.kernel.org/r/20190829163443.899-1-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reported-by: David Hildenbrand <david@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit ac311a14c6 ("oom: decouple mems_allowed from
oom_unkillable_task") changed has_intersects_mems_allowed() to
oom_cpuset_eligible(), but didn't change the comment.
Link: http://lkml.kernel.org/r/1566959929-10638-1-git-send-email-wang.yi59@zte.com.cn
Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For an OOM event: print oom_score_adj value for the OOM Killed process to
document what the oom score adjust value was at the time the process was
OOM Killed. The adjustment value can be set by user code and it affects
the resulting oom_score so it is used to influence kill process selection.
When eligible tasks are not printed (sysctl oom_dump_tasks = 0) printing
this value is the only documentation of the value for the process being
killed. Having this value on the Killed process message is useful to
document if a miscconfiguration occurred or to confirm that the
oom_score_adj configuration applies as expected.
An example which illustates both misconfiguration and validation that the
oom_score_adj was applied as expected is:
Aug 14 23:00:02 testserver kernel: Out of memory: Killed process 2692
(systemd-udevd) total-vm:1056800kB, anon-rss:1052760kB, file-rss:4kB,
shmem-rss:0kB pgtables:22kB oom_score_adj:1000
The systemd-udevd is a critical system application that should have an
oom_score_adj of -1000. It was miconfigured to have a adjustment of 1000
making it a highly favored OOM kill target process. The output documents
both the misconfiguration and the fact that the process was correctly
targeted by OOM due to the miconfiguration. This can be quite helpful for
triage and problem determination.
The addition of the pgtables_bytes shows page table usage by the process
and is a useful measure of the memory size of the process.
Link: http://lkml.kernel.org/r/20190822173157.1569-1-echron@arista.com
Signed-off-by: Edward Chron <echron@arista.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the event of an oom kill, useful information about the killed process
is printed to dmesg. Users, especially system administrators, will find
it useful to immediately see the UID of the process.
We already print uid when dumping eligible tasks so it is not overly hard
to find that information in the oom report. However this information is
unavailable when dumping of eligible tasks is disabled.
In the following example, abuse_the_ram is the name of a program that
attempts to iteratively allocate all available memory until it is stopped
by force.
Current message:
Out of memory: Killed process 35389 (abuse_the_ram)
total-vm:133718232kB, anon-rss:129624980kB, file-rss:0kB,
shmem-rss:0kB
Patched message:
Out of memory: Killed process 2739 (abuse_the_ram),
total-vm:133880028kB, anon-rss:129754836kB, file-rss:0kB,
shmem-rss:0kB, UID:0
[akpm@linux-foundation.org: s/UID %d/UID:%u/ in printk]
Link: http://lkml.kernel.org/r/1560362273-534-1-git-send-email-jsavitz@redhat.com
Signed-off-by: Joel Savitz <jsavitz@redhat.com>
Suggested-by: David Rientjes <rientjes@google.com>
Acked-by: Rafael Aquini <aquini@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1) task_nodes = cpuset_mems_allowed(current);
-> cpuset_mems_allowed() guaranteed to return some non-empty
subset of node_states[N_MEMORY].
2) nodes_and(*new, *new, task_nodes);
-> after nodes_and(), the 'new' should be empty or appropriate
nodemask(online node and with memory).
After 1) and 2), we could remove unnecessary check whether the 'new'
AND node_states[N_MEMORY] is empty.
Link: http://lkml.kernel.org/r/20190806023634.55356-1-wangkefeng.wang@huawei.com
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
total_{migrate,free}_scanned will be added to COMPACTMIGRATE_SCANNED and
COMPACTFREE_SCANNED in compact_zone(). We should clear them before
scanning a new zone. In the proc triggered compaction, we forgot clearing
them.
[laoar.shao@gmail.com: introduce a helper compact_zone_counters_init()]
Link: http://lkml.kernel.org/r/1563869295-25748-1-git-send-email-laoar.shao@gmail.com
[akpm@linux-foundation.org: expand compact_zone_counters_init() into its single callsite, per mhocko]
[vbabka@suse.cz: squash compact_zone() list_head init as well]
Link: http://lkml.kernel.org/r/1fb6f7da-f776-9e42-22f8-bbb79b030b98@suse.cz
[akpm@linux-foundation.org: kcompactd_do_work(): avoid unnecessary initialization of cc.zone]
Link: http://lkml.kernel.org/r/1563789275-9639-1-git-send-email-laoar.shao@gmail.com
Fixes: 7f354a548d ("mm, compaction: add vmstats for kcompactd work")
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Yafang Shao <shaoyafang@didiglobal.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently there is a leak in init_z3fold_page() -- it allocates handles
from kmem cache even for headless pages, but then they are never used and
never freed, so eventually kmem cache may get exhausted. This patch
provides a fix for that.
Link: http://lkml.kernel.org/r/20190917185352.44cf285d3ebd9e64548de5de@gmail.com
Signed-off-by: Vitaly Wool <vitalywool@gmail.com>
Reported-by: Markus Linnala <markus.linnala@gmail.com>
Tested-by: Markus Linnala <markus.linnala@gmail.com>
Cc: Dan Streetman <ddstreet@ieee.org>
Cc: Henry Burns <henrywolfeburns@gmail.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When compiling a kernel with W=1, there are several of those warnings due
to arm64 overriding a field on purpose. Just disable those warnings for
both GCC and Clang of this file, so it will help dig "gems" hidden in the
W=1 warnings by reducing some noises.
mm/init-mm.c:39:2: warning: initializer overrides prior initialization
of this subobject [-Winitializer-overrides]
INIT_MM_CONTEXT(init_mm)
^~~~~~~~~~~~~~~~~~~~~~~~
./arch/arm64/include/asm/mmu.h:133:9: note: expanded from macro
'INIT_MM_CONTEXT'
.pgd = init_pg_dir,
^~~~~~~~~~~
mm/init-mm.c:30:10: note: previous initialization is here
.pgd = swapper_pg_dir,
^~~~~~~~~~~~~~
Note: there is a side project trying to support explicitly allowing
specific initializer overrides in Clang, but there is no guarantee it
will happen or not.
https://github.com/ClangBuiltLinux/linux/issues/639
Link: http://lkml.kernel.org/r/1566920867-27453-1-git-send-email-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Replace open-coded bitmap array initialization of init_mm.cpu_bitmask with
neat CPU_BITS_NONE macro.
And, since init_mm.cpu_bitmask is statically set to zero, there is no way
to clear it again in start_kernel().
Link: http://lkml.kernel.org/r/1565703815-8584-1-git-send-email-rppt@linux.ibm.com
Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
If !area->pages statement is true where memory allocation fails, area is
freed.
In this case 'area->pages = pages' should not executed. So move
'area->pages = pages' after if statement.
[akpm@linux-foundation.org: give area->pages the same treatment]
Link: http://lkml.kernel.org/r/20190830035716.GA190684@LGEARND20B15
Signed-off-by: Austin Kim <austindh.kim@gmail.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Roman Penyaev <rpenyaev@suse.de>
Cc: Rick Edgecombe <rick.p.edgecombe@intel.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Objective
---------
The current implementation of struct vmap_area wasted space.
After applying this commit, sizeof(struct vmap_area) has been
reduced from 11 words to 8 words.
Description
-----------
1) Pack "subtree_max_size", "vm" and "purge_list". This is no problem
because
A) "subtree_max_size" is only used when vmap_area is in "free" tree
B) "vm" is only used when vmap_area is in "busy" tree
C) "purge_list" is only used when vmap_area is in vmap_purge_list
2) Eliminate "flags".
;Since only one flag VM_VM_AREA is being used, and the same thing can be
done by judging whether "vm" is NULL, then the "flags" can be eliminated.
Link: http://lkml.kernel.org/r/20190716152656.12255-3-lpf.vector@gmail.com
Signed-off-by: Pengfei Li <lpf.vector@gmail.com>
Suggested-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The busy tree can be quite big, even though the area is freed or unmapped
it still stays there until "purge" logic removes it.
1) Optimize and reduce the size of "busy" tree by removing a node from
it right away as soon as user triggers free paths. It is possible to
do so, because the allocation is done using another augmented tree.
The vmalloc test driver shows the difference, for example the
"fix_size_alloc_test" is ~11% better comparing with default configuration:
sudo ./test_vmalloc.sh performance
<default>
Summary: fix_size_alloc_test loops: 1000000 avg: 993985 usec
Summary: full_fit_alloc_test loops: 1000000 avg: 973554 usec
Summary: long_busy_list_alloc_test loops: 1000000 avg: 12617652 usec
<default>
<this patch>
Summary: fix_size_alloc_test loops: 1000000 avg: 882263 usec
Summary: full_fit_alloc_test loops: 1000000 avg: 973407 usec
Summary: long_busy_list_alloc_test loops: 1000000 avg: 12593929 usec
<this patch>
2) Since the busy tree now contains allocated areas only and does not
interfere with lazily free nodes, introduce the new function
show_purge_info() that dumps "unpurged" areas that is propagated
through "/proc/vmallocinfo".
3) Eliminate VM_LAZY_FREE flag.
Link: http://lkml.kernel.org/r/20190716152656.12255-2-lpf.vector@gmail.com
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Signed-off-by: Pengfei Li <lpf.vector@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Uladzislau Rezki <urezki@gmail.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Oleksiy Avramchenko <oleksiy.avramchenko@sonymobile.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There is no possibility for memmap to be NULL in the current codebase.
This check was added in commit 95a4774d05 ("memory-hotplug: update
mce_bad_pages when removing the memory") where memmap was originally
inited to NULL, and only conditionally given a value.
The code that could have passed a NULL has been removed by commit
ba72b4c8cf ("mm/sparsemem: support sub-section hotplug"), so there is no
longer a possibility that memmap can be NULL.
Link: http://lkml.kernel.org/r/20190829035151.20975-1-alastair@d-silva.org
Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Qian Cai <cai@lca.pw>
Cc: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use the function written to do it instead.
Link: http://lkml.kernel.org/r/20190827053656.32191-2-alastair@au1.ibm.com
Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Mike Rapoport <rppt@linux.ibm.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
__pfn_to_section is defined as __nr_to_section(pfn_to_section_nr(pfn)).
Since we already get section_nr, it is not necessary to get mem_section
from start_pfn. By doing so, we reduce one redundant operation.
Link: http://lkml.kernel.org/r/20190809010242.29797-1-richardw.yang@linux.intel.com
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Tested-by: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: David Hildenbrand <david@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The size argument passed into sparse_buffer_alloc() has already been
aligned with PAGE_SIZE or PMD_SIZE.
If the size after aligned is not power of 2 (e.g. 0x480000), the
PTR_ALIGN() will return wrong value. Use roundup to round sparsemap_buf
up to next multiple of size.
Link: http://lkml.kernel.org/r/20190705114826.28586-1-lecopzer.chen@mediatek.com
Signed-off-by: Lecopzer Chen <lecopzer.chen@mediatek.com>
Signed-off-by: Mark-PK Tsai <Mark-PK.Tsai@mediatek.com>
Cc: YJ Chiang <yj.chiang@mediatek.com>
Cc: Lecopzer Chen <lecopzer.chen@mediatek.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
sparse_buffer_alloc(xsize) gets the size of memory from sparsemap_buf
after being aligned with the size. However, the size is at least
PAGE_ALIGN(sizeof(struct page) * PAGES_PER_SECTION) and usually larger
than PAGE_SIZE.
Also, sparse_buffer_fini() only frees memory between sparsemap_buf and
sparsemap_buf_end, since sparsemap_buf may be changed by PTR_ALIGN()
first, the aligned space before sparsemap_buf is wasted and no one will
touch it.
In our ARM32 platform (without SPARSEMEM_VMEMMAP)
Sparse_buffer_init
Reserve d359c000 - d3e9c000 (9M)
Sparse_buffer_alloc
Alloc d3a00000 - d3E80000 (4.5M)
Sparse_buffer_fini
Free d3e80000 - d3e9c000 (~=100k)
The reserved memory between d359c000 - d3a00000 (~=4.4M) is unfreed.
In ARM64 platform (with SPARSEMEM_VMEMMAP)
sparse_buffer_init
Reserve ffffffc07d623000 - ffffffc07f623000 (32M)
Sparse_buffer_alloc
Alloc ffffffc07d800000 - ffffffc07f600000 (30M)
Sparse_buffer_fini
Free ffffffc07f600000 - ffffffc07f623000 (140K)
The reserved memory between ffffffc07d623000 - ffffffc07d800000
(~=1.9M) is unfreed.
Let's explicit free redundant aligned memory.
[arnd@arndb.de: mark sparse_buffer_free as __meminit]
Link: http://lkml.kernel.org/r/20190709185528.3251709-1-arnd@arndb.de
Link: http://lkml.kernel.org/r/20190705114730.28534-1-lecopzer.chen@mediatek.com
Signed-off-by: Lecopzer Chen <lecopzer.chen@mediatek.com>
Signed-off-by: Mark-PK Tsai <Mark-PK.Tsai@mediatek.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: YJ Chiang <yj.chiang@mediatek.com>
Cc: Lecopzer Chen <lecopzer.chen@mediatek.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
walk_system_ram_range() will fail with -EINVAL in case
online_pages_range() was never called (== no resource applicable in the
range). Otherwise, we will always call online_pages_range() with nr_pages
> 0 and, therefore, have online_pages > 0.
Remove that special handling.
Link: http://lkml.kernel.org/r/20190814154109.3448-6-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Nadav Amit <namit@vmware.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit a9cd410a3d ("mm/page_alloc.c: memory hotplug: free pages as
higher order") assumed that any PFN we get via memory resources is aligned
to to MAX_ORDER - 1, I am not convinced that is always true. Let's play
safe, check the alignment and fallback to single pages.
akpm: warn in this situation so we get to find out if and why this ever
occurs.
[akpm@linux-foundation.org: add WARN_ON_ONCE()]
Link: http://lkml.kernel.org/r/20190814154109.3448-5-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Nadav Amit <namit@vmware.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
online_pages always corresponds to nr_pages. Simplify the code, getting
rid of online_pages_blocks(). Add some comments.
Link: http://lkml.kernel.org/r/20190814154109.3448-4-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Nadav Amit <namit@vmware.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
move_pfn_range_to_zone() will set all pages to PG_reserved via
memmap_init_zone(). The only way a page could no longer be reserved would
be if a MEM_GOING_ONLINE notifier would clear PG_reserved - which is not
done (the online_page callback is used for that purpose by e.g., Hyper-V
instead). walk_system_ram_range() will never call online_pages_range()
with duplicate PFNs, so drop the PageReserved() check.
This seems to be a leftover from ancient times where the memmap was
initialized when adding memory and we wanted to check for already onlined
memory.
Link: http://lkml.kernel.org/r/20190814154109.3448-3-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Arun KS <arunks@codeaurora.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Nadav Amit <namit@vmware.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When offlining a node in try_offline_node(), pgdat is not released. So
that pgdat could be reused in hotadd_new_pgdat(). While we reallocate
pgdat->per_cpu_nodestats if this pgdat is reused.
This patch prevents the memory leak by just allocating per_cpu_nodestats
when it is a new pgdat.
Link: http://lkml.kernel.org/r/20190813020608.10194-1-richardw.yang@linux.intel.com
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Oscar Salvador <OSalvador@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Each memory block spans the same amount of sections/pages/bytes. The size
is determined before the first memory block is created. No need to store
what we can easily calculate - and the calculations even look simpler now.
Michal brought up the idea of variable-sized memory blocks. However, if
we ever implement something like this, we will need an API compatibility
switch and reworks at various places (most code assumes a fixed memory
block size). So let's cleanup what we have right now.
While at it, fix the variable naming in register_mem_sect_under_node() -
we no longer talk about a single section.
Link: http://lkml.kernel.org/r/20190809110200.2746-1-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Let's remove this indirection. We need the zone in the caller either way,
so let's just detect it there. Add some documentation for
move_pfn_range_to_zone() instead.
[akpm@linux-foundation.org: restore newline, per David]
Link: http://lkml.kernel.org/r/20190724142324.3686-1-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Using %px to show the actual address in print_bad_pte()
to help us to debug issue.
Link: http://lkml.kernel.org/r/20190831011816.141002-1-wangkefeng.wang@huawei.com
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm: remove quicklist page table caches".
A while ago Nicholas proposed to remove quicklist page table caches [1].
I've rebased his patch on the curren upstream and switched ia64 and sh to
use generic versions of PTE allocation.
[1] https://lore.kernel.org/linux-mm/20190711030339.20892-1-npiggin@gmail.com
This patch (of 3):
Remove page table allocator "quicklists". These have been around for a
long time, but have not got much traction in the last decade and are only
used on ia64 and sh architectures.
The numbers in the initial commit look interesting but probably don't
apply anymore. If anybody wants to resurrect this it's in the git
history, but it's unhelpful to have this code and divergent allocator
behaviour for minor archs.
Also it might be better to instead make more general improvements to page
allocator if this is still so slow.
Link: http://lkml.kernel.org/r/1565250728-21721-2-git-send-email-rppt@linux.ibm.com
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In our testing (camera recording), Miguel and Wei found
unmap_page_range() takes above 6ms with preemption disabled easily.
When I see that, the reason is it holds page table spinlock during
entire 512 page operation in a PMD. 6.2ms is never trivial for user
experince if RT task couldn't run in the time because it could make
frame drop or glitch audio problem.
I had a time to benchmark it via adding some trace_printk hooks between
pte_offset_map_lock and pte_unmap_unlock in zap_pte_range. The testing
device is 2018 premium mobile device.
I can get 2ms delay rather easily to release 2M(ie, 512 pages) when the
task runs on little core even though it doesn't have any IPI and LRU
lock contention. It's already too heavy.
If I remove activate_page, 35-40% overhead of zap_pte_range is gone so
most of overhead(about 0.7ms) comes from activate_page via
mark_page_accessed. Thus, if there are LRU contention, that 0.7ms could
accumulate up to several ms.
So this patch adds a check for need_resched() in the loop, and a
preemption point if necessary.
Link: http://lkml.kernel.org/r/20190731061440.GC155569@google.com
Signed-off-by: Minchan Kim <minchan@kernel.org>
Reported-by: Miguel de Dios <migueldedios@google.com>
Reported-by: Wei Wang <wvw@google.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since ptent will not be changed after previous assignment of entry, it is
not necessary to do the assignment again.
Link: http://lkml.kernel.org/r/20190708082740.21111-1-richardw.yang@linux.intel.com
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Acked-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Will Deacon <will@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[11~From: John Hubbard <jhubbard@nvidia.com>
Subject: mm/gup: add make_dirty arg to put_user_pages_dirty_lock()
Patch series "mm/gup: add make_dirty arg to put_user_pages_dirty_lock()",
v3.
There are about 50+ patches in my tree [2], and I'll be sending out the
remaining ones in a few more groups:
* The block/bio related changes (Jerome mostly wrote those, but I've had
to move stuff around extensively, and add a little code)
* mm/ changes
* other subsystem patches
* an RFC that shows the current state of the tracking patch set. That
can only be applied after all call sites are converted, but it's good to
get an early look at it.
This is part a tree-wide conversion, as described in fc1d8e7cca ("mm:
introduce put_user_page*(), placeholder versions").
This patch (of 3):
Provide more capable variation of put_user_pages_dirty_lock(), and delete
put_user_pages_dirty(). This is based on the following:
1. Lots of call sites become simpler if a bool is passed into
put_user_page*(), instead of making the call site choose which
put_user_page*() variant to call.
2. Christoph Hellwig's observation that set_page_dirty_lock() is
usually correct, and set_page_dirty() is usually a bug, or at least
questionable, within a put_user_page*() calling chain.
This leads to the following API choices:
* put_user_pages_dirty_lock(page, npages, make_dirty)
* There is no put_user_pages_dirty(). You have to
hand code that, in the rare case that it's
required.
[jhubbard@nvidia.com: remove unused variable in siw_free_plist()]
Link: http://lkml.kernel.org/r/20190729074306.10368-1-jhubbard@nvidia.com
Link: http://lkml.kernel.org/r/20190724044537.10458-2-jhubbard@nvidia.com
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Jan Kara <jack@suse.cz>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
One of our services observed a high rate of cgroup OOM kills in the
presence of large amounts of clean cache. Debugging showed that the
culprit is the shared cgroup iteration in page reclaim.
Under high allocation concurrency, multiple threads enter reclaim at the
same time. Fearing overreclaim when we first switched from the single
global LRU to cgrouped LRU lists, we introduced a shared iteration state
for reclaim invocations - whether 1 or 20 reclaimers are active
concurrently, we only walk the cgroup tree once: the 1st reclaimer
reclaims the first cgroup, the second the second one etc. With more
reclaimers than cgroups, we start another walk from the top.
This sounded reasonable at the time, but the problem is that reclaim
concurrency doesn't scale with allocation concurrency. As reclaim
concurrency increases, the amount of memory individual reclaimers get to
scan gets smaller and smaller. Individual reclaimers may only see one
cgroup per cycle, and that may not have much reclaimable memory. We see
individual reclaimers declare OOM when there is plenty of reclaimable
memory available in cgroups they didn't visit.
This patch does away with the shared iterator, and every reclaimer is
allowed to scan the full cgroup tree and see all of reclaimable memory,
just like it would on a non-cgrouped system. This way, when OOM is
declared, we know that the reclaimer actually had a chance.
To still maintain fairness in reclaim pressure, disallow cgroup reclaim
from bailing out of the tree walk early. Kswapd and regular direct
reclaim already don't bail, so it's not clear why limit reclaim would have
to, especially since it only walks subtrees to begin with.
This change completely eliminates the OOM kills on our service, while
showing no signs of overreclaim - no increased scan rates, %sys time, or
abrupt free memory spikes. I tested across 100 machines that have 64G of
RAM and host about 300 cgroups each.
[ It's possible overreclaim never was a *practical* issue to begin
with - it was simply a concern we had on the mailing lists at the
time, with no real data to back it up. But we have also added more
bail-out conditions deeper inside reclaim (e.g. the proportional
exit in shrink_node_memcg) since. Regardless, now we have data that
suggests full walks are more reliable and scale just fine. ]
Link: http://lkml.kernel.org/r/20190812192316.13615-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Roman Gushchin <guro@fb.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 72f0184c8a ("mm, memcg: remove hotplug locking from try_charge")
introduced css_tryget()/css_put() calls in drain_all_stock(), which are
supposed to protect the target memory cgroup from being released during
the mem_cgroup_is_descendant() call.
However, it's not completely safe. In theory, memcg can go away between
reading stock->cached pointer and calling css_tryget().
This can happen if drain_all_stock() races with drain_local_stock()
performed on the remote cpu as a result of a work, scheduled by the
previous invocation of drain_all_stock().
The race is a bit theoretical and there are few chances to trigger it, but
the current code looks a bit confusing, so it makes sense to fix it
anyway. The code looks like as if css_tryget() and css_put() are used to
protect stocks drainage. It's not necessary because stocked pages are
holding references to the cached cgroup. And it obviously won't work for
works, scheduled on other cpus.
So, let's read the stock->cached pointer and evaluate the memory cgroup
inside a rcu read section, and get rid of css_tryget()/css_put() calls.
Link: http://lkml.kernel.org/r/20190802192241.3253165-1-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We're trying to use memory.high to limit workloads, but have found that
containment can frequently fail completely and cause OOM situations
outside of the cgroup. This happens especially with swap space -- either
when none is configured, or swap is full. These failures often also don't
have enough warning to allow one to react, whether for a human or for a
daemon monitoring PSI.
Here is output from a simple program showing how long it takes in usec
(column 2) to allocate a megabyte of anonymous memory (column 1) when a
cgroup is already beyond its memory high setting, and no swap is
available:
[root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \
> --wait -t timeout 300 /root/mdf
[...]
95 1035
96 1038
97 1000
98 1036
99 1048
100 1590
101 1968
102 1776
103 1863
104 1757
105 1921
106 1893
107 1760
108 1748
109 1843
110 1716
111 1924
112 1776
113 1831
114 1766
115 1836
116 1588
117 1912
118 1802
119 1857
120 1731
[...]
[System OOM in 2-3 seconds]
The delay does go up extremely marginally past the 100MB memory.high
threshold, as now we spend time scanning before returning to usermode, but
it's nowhere near enough to contain growth. It also doesn't get worse the
more pages you have, since it only considers nr_pages.
The current situation goes against both the expectations of users of
memory.high, and our intentions as cgroup v2 developers. In
cgroup-v2.txt, we claim that we will throttle and only under "extreme
conditions" will memory.high protection be breached. Likewise, cgroup v2
users generally also expect that memory.high should throttle workloads as
they exceed their high threshold. However, as seen above, this isn't
always how it works in practice -- even on banal setups like those with no
swap, or where swap has become exhausted, we can end up with memory.high
being breached and us having no weapons left in our arsenal to combat
runaway growth with, since reclaim is futile.
It's also hard for system monitoring software or users to tell how bad the
situation is, as "high" events for the memcg may in some cases be benign,
and in others be catastrophic. The current status quo is that we fail
containment in a way that doesn't provide any advance warning that things
are about to go horribly wrong (for example, we are about to invoke the
kernel OOM killer).
This patch introduces explicit throttling when reclaim is failing to keep
memcg size contained at the memory.high setting. It does so by applying
an exponential delay curve derived from the memcg's overage compared to
memory.high. In the normal case where the memcg is either below or only
marginally over its memory.high setting, no throttling will be performed.
This composes well with system health monitoring and remediation, as these
allocator delays are factored into PSI's memory pressure calculations.
This both creates a mechanism system administrators or applications
consuming the PSI interface to trivially see that the memcg in question is
struggling and use that to make more reasonable decisions, and permits
them enough time to act. Either of these can act with significantly more
nuance than that we can provide using the system OOM killer.
This is a similar idea to memory.oom_control in cgroup v1 which would put
the cgroup to sleep if the threshold was violated, but it's also
significantly improved as it results in visible memory pressure, and also
doesn't schedule indefinitely, which previously made tracing and other
introspection difficult (ie. it's clamped at 2*HZ per allocation through
MEMCG_MAX_HIGH_DELAY_JIFFIES).
Contrast the previous results with a kernel with this patch:
[root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \
> --wait -t timeout 300 /root/mdf
[...]
95 1002
96 1000
97 1002
98 1003
99 1000
100 1043
101 84724
102 330628
103 610511
104 1016265
105 1503969
106 2391692
107 2872061
108 3248003
109 4791904
110 5759832
111 6912509
112 8127818
113 9472203
114 12287622
115 12480079
116 14144008
117 15808029
118 16384500
119 16383242
120 16384979
[...]
As you can see, in the normal case, memory allocation takes around 1000
usec. However, as we exceed our memory.high, things start to increase
exponentially, but fairly leniently at first. Our first megabyte over
memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the
next is almost an entire second. This gets worse until we reach our
eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte.
However, this is still making forward progress, so permits tracing or
further analysis with programs like GDB.
We use an exponential curve for our delay penalty for a few reasons:
1. We run mem_cgroup_handle_over_high to potentially do reclaim after
we've already performed allocations, which means that temporarily
going over memory.high by a small amount may be perfectly legitimate,
even for compliant workloads. We don't want to unduly penalise such
cases.
2. An exponential curve (as opposed to a static or linear delay) allows
ramping up memory pressure stats more gradually, which can be useful
to work out that you have set memory.high too low, without destroying
application performance entirely.
This patch expands on earlier work by Johannes Weiner. Thanks!
[akpm@linux-foundation.org: fix max() warning]
[akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit]
[akpm@linux-foundation.org: fix it even more]
[chris@chrisdown.name: fix 64-bit divide even more]
Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name
Signed-off-by: Chris Down <chris@chrisdown.name>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Roman Gushchin <guro@fb.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Transparent Huge Pages are currently stored in i_pages as pointers to
consecutive subpages. This patch changes that to storing consecutive
pointers to the head page in preparation for storing huge pages more
efficiently in i_pages.
Large parts of this are "inspired" by Kirill's patch
https://lore.kernel.org/lkml/20170126115819.58875-2-kirill.shutemov@linux.intel.com/
Kirill and Huang Ying contributed several fixes.
[willy@infradead.org: use compound_nr, squish uninit-var warning]
Link: http://lkml.kernel.org/r/20190731210400.7419-1-willy@infradead.org
Signed-off-by: Matthew Wilcox <willy@infradead.org>
Acked-by: Jan Kara <jack@suse.cz>
Reviewed-by: Kirill Shutemov <kirill@shutemov.name>
Reviewed-by: Song Liu <songliubraving@fb.com>
Tested-by: Song Liu <songliubraving@fb.com>
Tested-by: William Kucharski <william.kucharski@oracle.com>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Tested-by: Qian Cai <cai@lca.pw>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Song Liu <songliubraving@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This actually checks that writeback is needed or in progress.
Link: http://lkml.kernel.org/r/156378817069.1087.1302816672037672488.stgit@buzz
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Functions like filemap_write_and_wait_range() should do nothing if inode
has no dirty pages or pages currently under writeback. But they anyway
construct struct writeback_control and this does some atomic operations if
CONFIG_CGROUP_WRITEBACK=y - on fast path it locks inode->i_lock and
updates state of writeback ownership, on slow path might be more work.
Current this path is safely avoided only when inode mapping has no pages.
For example generic_file_read_iter() calls filemap_write_and_wait_range()
at each O_DIRECT read - pretty hot path.
This patch skips starting new writeback if mapping has no dirty tags set.
If writeback is already in progress filemap_write_and_wait_range() will
wait for it.
Link: http://lkml.kernel.org/r/156378816804.1087.8607636317907921438.stgit@buzz
Signed-off-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The debug_pagealloc functionality is useful to catch buggy page allocator
users that cause e.g. use after free or double free. When page
inconsistency is detected, debugging is often simpler by knowing the call
stack of process that last allocated and freed the page. When page_owner
is also enabled, we record the allocation stack trace, but not freeing.
This patch therefore adds recording of freeing process stack trace to page
owner info, if both page_owner and debug_pagealloc are configured and
enabled. With only page_owner enabled, this info is not useful for the
memory leak debugging use case. dump_page() is adjusted to print the
info. An example result of calling __free_pages() twice may look like
this (note the page last free stack trace):
BUG: Bad page state in process bash pfn:13d8f8
page:ffffc31984f63e00 refcount:-1 mapcount:0 mapping:0000000000000000 index:0x0
flags: 0x1affff800000000()
raw: 01affff800000000 dead000000000100 dead000000000122 0000000000000000
raw: 0000000000000000 0000000000000000 ffffffffffffffff 0000000000000000
page dumped because: nonzero _refcount
page_owner tracks the page as freed
page last allocated via order 0, migratetype Unmovable, gfp_mask 0xcc0(GFP_KERNEL)
prep_new_page+0x143/0x150
get_page_from_freelist+0x289/0x380
__alloc_pages_nodemask+0x13c/0x2d0
khugepaged+0x6e/0xc10
kthread+0xf9/0x130
ret_from_fork+0x3a/0x50
page last free stack trace:
free_pcp_prepare+0x134/0x1e0
free_unref_page+0x18/0x90
khugepaged+0x7b/0xc10
kthread+0xf9/0x130
ret_from_fork+0x3a/0x50
Modules linked in:
CPU: 3 PID: 271 Comm: bash Not tainted 5.3.0-rc4-2.g07a1a73-default+ #57
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.1-0-ga5cab58-prebuilt.qemu.org 04/01/2014
Call Trace:
dump_stack+0x85/0xc0
bad_page.cold+0xba/0xbf
rmqueue_pcplist.isra.0+0x6c5/0x6d0
rmqueue+0x2d/0x810
get_page_from_freelist+0x191/0x380
__alloc_pages_nodemask+0x13c/0x2d0
__get_free_pages+0xd/0x30
__pud_alloc+0x2c/0x110
copy_page_range+0x4f9/0x630
dup_mmap+0x362/0x480
dup_mm+0x68/0x110
copy_process+0x19e1/0x1b40
_do_fork+0x73/0x310
__x64_sys_clone+0x75/0x80
do_syscall_64+0x6e/0x1e0
entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x7f10af854a10
...
Link: http://lkml.kernel.org/r/20190820131828.22684-5-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For debugging purposes it might be useful to keep the owner info even
after page has been freed, and include it in e.g. dump_page() when
detecting a bad page state. For that, change the PAGE_EXT_OWNER flag
meaning to "page owner info has been set at least once" and add new
PAGE_EXT_OWNER_ACTIVE for tracking whether page is supposed to be
currently tracked allocated or free. Adjust dump_page() accordingly,
distinguishing free and allocated pages. In the page_owner debugfs file,
keep printing only allocated pages so that existing scripts are not
confused, and also because free pages are irrelevant for the memory
statistics or leak detection that's the typical use case of the file,
anyway.
Link: http://lkml.kernel.org/r/20190820131828.22684-4-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "debug_pagealloc improvements through page_owner", v2.
The debug_pagealloc functionality serves a similar purpose on the page
allocator level that slub_debug does on the kmalloc level, which is to
detect bad users. One notable feature that slub_debug has is storing
stack traces of who last allocated and freed the object. On page level we
track allocations via page_owner, but that info is discarded when freeing,
and we don't track freeing at all. This series improves those aspects.
With both debug_pagealloc and page_owner enabled, we can then get bug
reports such as the example in Patch 4.
SLUB debug tracking additionally stores cpu, pid and timestamp. This could
be added later, if deemed useful enough to justify the additional page_ext
structure size.
This patch (of 3):
Currently, page owner info is only recorded for the first page of a
high-order allocation, and copied to tail pages in the event of a split
page. With the plan to keep previous owner info after freeing the page,
it would be benefical to record page owner for each subpage upon
allocation. This increases the overhead for high orders, but that should
be acceptable for a debugging option.
The order stored for each subpage is the order of the whole allocation.
This makes it possible to calculate the "head" pfn and to recognize "tail"
pages (quoted because not all high-order allocations are compound pages
with true head and tail pages). When reading the page_owner debugfs file,
keep skipping the "tail" pages so that stats gathered by existing scripts
don't get inflated.
Link: http://lkml.kernel.org/r/20190820131828.22684-3-vbabka@suse.cz
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is a cleanup patch that replaces two historical uses of
list_move_tail() with relatively recent add_page_to_lru_list_tail().
Link: http://lkml.kernel.org/r/20190716212436.7137-1-yuzhao@google.com
Signed-off-by: Yu Zhao <yuzhao@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
