It was reported that on some of our machines containers were restarted
with OOM symptoms without an obvious reason. Despite there were almost no
memory pressure and plenty of page cache, MEMCG_OOM event was raised
occasionally, causing the container management software to think, that OOM
has happened. However, no tasks have been killed.
The following investigation showed that the problem is caused by a failing
attempt to charge a high-order page. In such case, the OOM killer is
never invoked. As shown below, it can happen under conditions, which are
very far from a real OOM: e.g. there is plenty of clean page cache and no
memory pressure.
There is no sense in raising an OOM event in this case, as it might
confuse a user and lead to wrong and excessive actions (e.g. restart the
workload, as in my case).
Let's look at the charging path in try_charge(). If the memory usage is
about memory.max, which is absolutely natural for most memory cgroups, we
try to reclaim some pages. Even if we were able to reclaim enough memory
for the allocation, the following check can fail due to a race with
another concurrent allocation:
if (mem_cgroup_margin(mem_over_limit) >= nr_pages)
goto retry;
For regular pages the following condition will save us from triggering
the OOM:
if (nr_reclaimed && nr_pages <= (1 << PAGE_ALLOC_COSTLY_ORDER))
goto retry;
But for high-order allocation this condition will intentionally fail. The
reason behind is that we'll likely fall to regular pages anyway, so it's
ok and even preferred to return ENOMEM.
In this case the idea of raising MEMCG_OOM looks dubious.
Fix this by moving MEMCG_OOM raising to mem_cgroup_oom() after allocation
order check, so that the event won't be raised for high order allocations.
This change doesn't affect regular pages allocation and charging.
Link: http://lkml.kernel.org/r/20181004214050.7417-1-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Michal Hocko <mhocko@kernel.org>
Acked-by: 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've recently seen a workload on XFS filesystems with a repeatable
deadlock between background writeback and a multi-process application
doing concurrent writes and fsyncs to a small range of a file.
range_cyclic
writeback Process 1 Process 2
xfs_vm_writepages
write_cache_pages
writeback_index = 2
cycled = 0
....
find page 2 dirty
lock Page 2
->writepage
page 2 writeback
page 2 clean
page 2 added to bio
no more pages
write()
locks page 1
dirties page 1
locks page 2
dirties page 1
fsync()
....
xfs_vm_writepages
write_cache_pages
start index 0
find page 1 towrite
lock Page 1
->writepage
page 1 writeback
page 1 clean
page 1 added to bio
find page 2 towrite
lock Page 2
page 2 is writeback
<blocks>
write()
locks page 1
dirties page 1
fsync()
....
xfs_vm_writepages
write_cache_pages
start index 0
!done && !cycled
sets index to 0, restarts lookup
find page 1 dirty
find page 1 towrite
lock Page 1
page 1 is writeback
<blocks>
lock Page 1
<blocks>
DEADLOCK because:
- process 1 needs page 2 writeback to complete to make
enough progress to issue IO pending for page 1
- writeback needs page 1 writeback to complete so process 2
can progress and unlock the page it is blocked on, then it
can issue the IO pending for page 2
- process 2 can't make progress until process 1 issues IO
for page 1
The underlying cause of the problem here is that range_cyclic writeback is
processing pages in descending index order as we hold higher index pages
in a structure controlled from above write_cache_pages(). The
write_cache_pages() caller needs to be able to submit these pages for IO
before write_cache_pages restarts writeback at mapping index 0 to avoid
wcp inverting the page lock/writeback wait order.
generic_writepages() is not susceptible to this bug as it has no private
context held across write_cache_pages() - filesystems using this
infrastructure always submit pages in ->writepage immediately and so there
is no problem with range_cyclic going back to mapping index 0.
However:
mpage_writepages() has a private bio context,
exofs_writepages() has page_collect
fuse_writepages() has fuse_fill_wb_data
nfs_writepages() has nfs_pageio_descriptor
xfs_vm_writepages() has xfs_writepage_ctx
All of these ->writepages implementations can hold pages under writeback
in their private structures until write_cache_pages() returns, and hence
they are all susceptible to this deadlock.
Also worth noting is that ext4 has it's own bastardised version of
write_cache_pages() and so it /may/ have an equivalent deadlock. I looked
at the code long enough to understand that it has a similar retry loop for
range_cyclic writeback reaching the end of the file and then promptly ran
away before my eyes bled too much. I'll leave it for the ext4 developers
to determine if their code is actually has this deadlock and how to fix it
if it has.
There's a few ways I can see avoid this deadlock. There's probably more,
but these are the first I've though of:
1. get rid of range_cyclic altogether
2. range_cyclic always stops at EOF, and we start again from
writeback index 0 on the next call into write_cache_pages()
2a. wcp also returns EAGAIN to ->writepages implementations to
indicate range cyclic has hit EOF. writepages implementations can
then flush the current context and call wpc again to continue. i.e.
lift the retry into the ->writepages implementation
3. range_cyclic uses trylock_page() rather than lock_page(), and it
skips pages it can't lock without blocking. It will already do this
for pages under writeback, so this seems like a no-brainer
3a. all non-WB_SYNC_ALL writeback uses trylock_page() to avoid
blocking as per pages under writeback.
I don't think #1 is an option - range_cyclic prevents frequently
dirtied lower file offset from starving background writeback of
rarely touched higher file offsets.
#2 is simple, and I don't think it will have any impact on
performance as going back to the start of the file implies an
immediate seek. We'll have exactly the same number of seeks if we
switch writeback to another inode, and then come back to this one
later and restart from index 0.
#2a is pretty much "status quo without the deadlock". Moving the
retry loop up into the wcp caller means we can issue IO on the
pending pages before calling wcp again, and so avoid locking or
waiting on pages in the wrong order. I'm not convinced we need to do
this given that we get the same thing from #2 on the next writeback
call from the writeback infrastructure.
#3 is really just a band-aid - it doesn't fix the access/wait
inversion problem, just prevents it from becoming a deadlock
situation. I'd prefer we fix the inversion, not sweep it under the
carpet like this.
#3a is really an optimisation that just so happens to include the
band-aid fix of #3.
So it seems that the simplest way to fix this issue is to implement
solution #2
Link: http://lkml.kernel.org/r/20181005054526.21507-1-david@fromorbit.com
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Jan Kara <jack@suse.de>
Cc: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Now we recycle the uffd servicing threads earlier than the lock threads.
It might happen that when the lock thread is still blocked at a pthread
mutex lock while the servicing thread has already quitted for the cpu so
the lock thread will be blocked forever and hang the test program. To fix
the possible race, recycle the lock threads first.
This never happens with current missing-only tests, but when I start to
run the write-protection tests (the feature is not yet posted upstream) it
happens every time of the run possibly because in that new test we'll need
to service two page faults for each lock operation.
Link: http://lkml.kernel.org/r/20180930074259.18229-4-peterx@redhat.com
Signed-off-by: Peter Xu <peterx@redhat.com>
Acked-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Zi Yan <zi.yan@cs.rutgers.edu>
Cc: "Kirill A . Shutemov" <kirill@shutemov.name>
Cc: Shaohua Li <shli@fb.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: "Dr . David Alan Gilbert" <dgilbert@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
brk might be used to shrink memory mapping too other than munmap(). So,
it may hold write mmap_sem for long time when shrinking large mapping, as
what commit ("mm: mmap: zap pages with read mmap_sem in munmap")
described.
The brk() will not manipulate vmas anymore after __do_munmap() call for
the mapping shrink use case. But, it may set mm->brk after __do_munmap(),
which needs hold write mmap_sem.
However, a simple trick can workaround this by setting mm->brk before
__do_munmap(). Then restore the original value if __do_munmap() fails.
With this trick, it is safe to downgrade to read mmap_sem.
So, the same optimization, which downgrades mmap_sem to read for zapping
pages, is also feasible and reasonable to this case.
The period of holding exclusive mmap_sem for shrinking large mapping would
be reduced significantly with this optimization.
[akpm@linux-foundation.org: tweak comment]
[yang.shi@linux.alibaba.com: fix unsigned compare against 0 issue]
Link: http://lkml.kernel.org/r/1538687672-17795-1-git-send-email-yang.shi@linux.alibaba.com
Link: http://lkml.kernel.org/r/1538067582-60038-2-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Cc: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Other than munmap, mremap might be used to shrink memory mapping too.
So, it may hold write mmap_sem for long time when shrinking large
mapping, as what commit ("mm: mmap: zap pages with read mmap_sem in
munmap") described.
The mremap() will not manipulate vmas anymore after __do_munmap() call for
the mapping shrink use case, so it is safe to downgrade to read mmap_sem.
So, the same optimization, which downgrades mmap_sem to read for zapping
pages, is also feasible and reasonable to this case.
The period of holding exclusive mmap_sem for shrinking large mapping
would be reduced significantly with this optimization.
MREMAP_FIXED and MREMAP_MAYMOVE are more complicated to adopt this
optimization since they need manipulate vmas after do_munmap(),
downgrading mmap_sem may create race window.
Simple mapping shrink is the low hanging fruit, and it may cover the
most cases of unmap with munmap together.
[akpm@linux-foundation.org: tweak comment]
[yang.shi@linux.alibaba.com: fix unsigned compare against 0 issue]
Link: http://lkml.kernel.org/r/1538687672-17795-2-git-send-email-yang.shi@linux.alibaba.com
Link: http://lkml.kernel.org/r/1538067582-60038-1-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Cc: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The ZONE_DEVICE pages were being initialized in two locations. One was
with the memory_hotplug lock held and another was outside of that lock.
The problem with this is that it was nearly doubling the memory
initialization time. Instead of doing this twice, once while holding a
global lock and once without, I am opting to defer the initialization to
the one outside of the lock. This allows us to avoid serializing the
overhead for memory init and we can instead focus on per-node init times.
One issue I encountered is that devm_memremap_pages and
hmm_devmmem_pages_create were initializing only the pgmap field the same
way. One wasn't initializing hmm_data, and the other was initializing it
to a poison value. Since this is something that is exposed to the driver
in the case of hmm I am opting for a third option and just initializing
hmm_data to 0 since this is going to be exposed to unknown third party
drivers.
[alexander.h.duyck@linux.intel.com: fix reference count for pgmap in devm_memremap_pages]
Link: http://lkml.kernel.org/r/20181008233404.1909.37302.stgit@localhost.localdomain
Link: http://lkml.kernel.org/r/20180925202053.3576.66039.stgit@localhost.localdomain
Signed-off-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Reviewed-by: Pavel Tatashin <pavel.tatashin@microsoft.com>
Tested-by: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It doesn't make much sense to use the atomic SetPageReserved at init time
when we are using memset to clear the memory and manipulating the page
flags via simple "&=" and "|=" operations in __init_single_page.
This patch adds a non-atomic version __SetPageReserved that can be used
during page init and shows about a 10% improvement in initialization times
on the systems I have available for testing. On those systems I saw
initialization times drop from around 35 seconds to around 32 seconds to
initialize a 3TB block of persistent memory. I believe the main advantage
of this is that it allows for more compiler optimization as the __set_bit
operation can be reordered whereas the atomic version cannot.
I tried adding a bit of documentation based on f1dd2cd13c ("mm,
memory_hotplug: do not associate hotadded memory to zones until online").
Ideally the reserved flag should be set earlier since there is a brief
window where the page is initialization via __init_single_page and we have
not set the PG_Reserved flag. I'm leaving that for a future patch set as
that will require a more significant refactor.
Link: http://lkml.kernel.org/r/20180925202018.3576.11607.stgit@localhost.localdomain
Signed-off-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Reviewed-by: Pavel Tatashin <pavel.tatashin@microsoft.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "Address issues slowing persistent memory initialization", v5.
The main thing this patch set achieves is that it allows us to initialize
each node worth of persistent memory independently. As a result we reduce
page init time by about 2 minutes because instead of taking 30 to 40
seconds per node and going through each node one at a time, we process all
4 nodes in parallel in the case of a 12TB persistent memory setup spread
evenly over 4 nodes.
This patch (of 3):
On systems with a large amount of memory it can take a significant amount
of time to initialize all of the page structs with the PAGE_POISON_PATTERN
value. I have seen it take over 2 minutes to initialize a system with
over 12TB of RAM.
In order to work around the issue I had to disable CONFIG_DEBUG_VM and
then the boot time returned to something much more reasonable as the
arch_add_memory call completed in milliseconds versus seconds. However in
doing that I had to disable all of the other VM debugging on the system.
In order to work around a kernel that might have CONFIG_DEBUG_VM enabled
on a system that has a large amount of memory I have added a new kernel
parameter named "vm_debug" that can be set to "-" in order to disable it.
Link: http://lkml.kernel.org/r/20180925201921.3576.84239.stgit@localhost.localdomain
Reviewed-by: Pavel Tatashin <pavel.tatashin@microsoft.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch, as the previous one, gets rid of the wrong if statements.
While at it, I realized that the comments are sometimes very confusing,
to say the least, and wrong.
For example:
___
zone_last = ZONE_MOVABLE;
/*
* check whether node_states[N_HIGH_MEMORY] will be changed
* If we try to offline the last present @nr_pages from the node,
* we can determind we will need to clear the node from
* node_states[N_HIGH_MEMORY].
*/
for (; zt <= zone_last; zt++)
present_pages += pgdat->node_zones[zt].present_pages;
if (nr_pages >= present_pages)
arg->status_change_nid = zone_to_nid(zone);
else
arg->status_change_nid = -1;
___
In case the node gets empry, it must be removed from N_MEMORY. We already
check N_HIGH_MEMORY a bit above within the CONFIG_HIGHMEM ifdef code. Not
to say that status_change_nid is for N_MEMORY, and not for N_HIGH_MEMORY.
So I re-wrote some of the comments to what I think is better.
[osalvador@suse.de: address feedback from Pavel]
Link: http://lkml.kernel.org/r/20180921132634.10103-5-osalvador@techadventures.net
Link: http://lkml.kernel.org/r/20180919100819.25518-6-osalvador@techadventures.net
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Pavel Tatashin <pavel.tatashin@microsoft.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: <yasu.isimatu@gmail.com>
Cc: Mathieu Malaterre <malat@debian.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm: zap pages with read mmap_sem in munmap for large
mapping", v11.
Background:
Recently, when we ran some vm scalability tests on machines with large memory,
we ran into a couple of mmap_sem scalability issues when unmapping large memory
space, please refer to https://lkml.org/lkml/2017/12/14/733 and
https://lkml.org/lkml/2018/2/20/576.
History:
Then akpm suggested to unmap large mapping section by section and drop mmap_sem
at a time to mitigate it (see https://lkml.org/lkml/2018/3/6/784).
V1 patch series was submitted to the mailing list per Andrew's suggestion
(see https://lkml.org/lkml/2018/3/20/786). Then I received a lot great
feedback and suggestions.
Then this topic was discussed on LSFMM summit 2018. In the summit, Michal
Hocko suggested (also in the v1 patches review) to try "two phases"
approach. Zapping pages with read mmap_sem, then doing via cleanup with
write mmap_sem (for discussion detail, see
https://lwn.net/Articles/753269/)
Approach:
Zapping pages is the most time consuming part, according to the suggestion from
Michal Hocko [1], zapping pages can be done with holding read mmap_sem, like
what MADV_DONTNEED does. Then re-acquire write mmap_sem to cleanup vmas.
But, we can't call MADV_DONTNEED directly, since there are two major drawbacks:
* The unexpected state from PF if it wins the race in the middle of munmap.
It may return zero page, instead of the content or SIGSEGV.
* Can't handle VM_LOCKED | VM_HUGETLB | VM_PFNMAP and uprobe mappings, which
is a showstopper from akpm
But, some part may need write mmap_sem, for example, vma splitting. So,
the design is as follows:
acquire write mmap_sem
lookup vmas (find and split vmas)
deal with special mappings
detach vmas
downgrade_write
zap pages
free page tables
release mmap_sem
The vm events with read mmap_sem may come in during page zapping, but
since vmas have been detached before, they, i.e. page fault, gup, etc,
will not be able to find valid vma, then just return SIGSEGV or -EFAULT as
expected.
If the vma has VM_HUGETLB | VM_PFNMAP, they are considered as special
mappings. They will be handled by falling back to regular do_munmap()
with exclusive mmap_sem held in this patch since they may update vm flags.
But, with the "detach vmas first" approach, the vmas have been detached
when vm flags are updated, so it sounds safe to update vm flags with read
mmap_sem for this specific case. So, VM_HUGETLB and VM_PFNMAP will be
handled by using the optimized path in the following separate patches for
bisectable sake.
Unmapping uprobe areas may need update mm flags (MMF_RECALC_UPROBES).
However it is fine to have false-positive MMF_RECALC_UPROBES according to
uprobes developer. So, uprobe unmap will not be handled by the regular
path.
With the "detach vmas first" approach we don't have to re-acquire mmap_sem
again to clean up vmas to avoid race window which might get the address
space changed since downgrade_write() doesn't release the lock to lead
regression, which simply downgrades to read lock.
And, since the lock acquire/release cost is managed to the minimum and
almost as same as before, the optimization could be extended to any size
of mapping without incurring significant penalty to small mappings.
For the time being, just do this in munmap syscall path. Other
vm_munmap() or do_munmap() call sites (i.e mmap, mremap, etc) remain
intact due to some implementation difficulties since they acquire write
mmap_sem from very beginning and hold it until the end, do_munmap() might
be called in the middle. But, the optimized do_munmap would like to be
called without mmap_sem held so that we can do the optimization. So, if
we want to do the similar optimization for mmap/mremap path, I'm afraid we
would have to redesign them. mremap might be called on very large area
depending on the usecases, the optimization to it will be considered in
the future.
This patch (of 3):
When running some mmap/munmap scalability tests with large memory (i.e.
> 300GB), the below hung task issue may happen occasionally.
INFO: task ps:14018 blocked for more than 120 seconds.
Tainted: G E 4.9.79-009.ali3000.alios7.x86_64 #1
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this
message.
ps D 0 14018 1 0x00000004
ffff885582f84000 ffff885e8682f000 ffff880972943000 ffff885ebf499bc0
ffff8828ee120000 ffffc900349bfca8 ffffffff817154d0 0000000000000040
00ffffff812f872a ffff885ebf499bc0 024000d000948300 ffff880972943000
Call Trace:
[<ffffffff817154d0>] ? __schedule+0x250/0x730
[<ffffffff817159e6>] schedule+0x36/0x80
[<ffffffff81718560>] rwsem_down_read_failed+0xf0/0x150
[<ffffffff81390a28>] call_rwsem_down_read_failed+0x18/0x30
[<ffffffff81717db0>] down_read+0x20/0x40
[<ffffffff812b9439>] proc_pid_cmdline_read+0xd9/0x4e0
[<ffffffff81253c95>] ? do_filp_open+0xa5/0x100
[<ffffffff81241d87>] __vfs_read+0x37/0x150
[<ffffffff812f824b>] ? security_file_permission+0x9b/0xc0
[<ffffffff81242266>] vfs_read+0x96/0x130
[<ffffffff812437b5>] SyS_read+0x55/0xc0
[<ffffffff8171a6da>] entry_SYSCALL_64_fastpath+0x1a/0xc5
It is because munmap holds mmap_sem exclusively from very beginning to all
the way down to the end, and doesn't release it in the middle. When
unmapping large mapping, it may take long time (take ~18 seconds to unmap
320GB mapping with every single page mapped on an idle machine).
Zapping pages is the most time consuming part, according to the suggestion
from Michal Hocko [1], zapping pages can be done with holding read
mmap_sem, like what MADV_DONTNEED does. Then re-acquire write mmap_sem to
cleanup vmas.
But, some part may need write mmap_sem, for example, vma splitting. So,
the design is as follows:
acquire write mmap_sem
lookup vmas (find and split vmas)
deal with special mappings
detach vmas
downgrade_write
zap pages
free page tables
release mmap_sem
The vm events with read mmap_sem may come in during page zapping, but
since vmas have been detached before, they, i.e. page fault, gup, etc,
will not be able to find valid vma, then just return SIGSEGV or -EFAULT as
expected.
If the vma has VM_HUGETLB | VM_PFNMAP, they are considered as special
mappings. They will be handled by without downgrading mmap_sem in this
patch since they may update vm flags.
But, with the "detach vmas first" approach, the vmas have been detached
when vm flags are updated, so it sounds safe to update vm flags with read
mmap_sem for this specific case. So, VM_HUGETLB and VM_PFNMAP will be
handled by using the optimized path in the following separate patches for
bisectable sake.
Unmapping uprobe areas may need update mm flags (MMF_RECALC_UPROBES).
However it is fine to have false-positive MMF_RECALC_UPROBES according to
uprobes developer.
With the "detach vmas first" approach we don't have to re-acquire mmap_sem
again to clean up vmas to avoid race window which might get the address
space changed since downgrade_write() doesn't release the lock to lead
regression, which simply downgrades to read lock.
And, since the lock acquire/release cost is managed to the minimum and
almost as same as before, the optimization could be extended to any size
of mapping without incurring significant penalty to small mappings.
For the time being, just do this in munmap syscall path. Other
vm_munmap() or do_munmap() call sites (i.e mmap, mremap, etc) remain
intact due to some implementation difficulties since they acquire write
mmap_sem from very beginning and hold it until the end, do_munmap() might
be called in the middle. But, the optimized do_munmap would like to be
called without mmap_sem held so that we can do the optimization. So, if
we want to do the similar optimization for mmap/mremap path, I'm afraid we
would have to redesign them. mremap might be called on very large area
depending on the usecases, the optimization to it will be considered in
the future.
With the patches, exclusive mmap_sem hold time when munmap a 80GB address
space on a machine with 32 cores of E5-2680 @ 2.70GHz dropped to us level
from second.
munmap_test-15002 [008] 594.380138: funcgraph_entry: |
__vm_munmap() {
munmap_test-15002 [008] 594.380146: funcgraph_entry: !2485684 us
| unmap_region();
munmap_test-15002 [008] 596.865836: funcgraph_exit: !2485692 us
| }
Here the execution time of unmap_region() is used to evaluate the time of
holding read mmap_sem, then the remaining time is used with holding
exclusive lock.
[1] https://lwn.net/Articles/753269/
Link: http://lkml.kernel.org/r/1537376621-51150-2-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>Suggested-by: Michal Hocko <mhocko@kernel.org>
Suggested-by: Kirill A. Shutemov <kirill@shutemov.name>
Suggested-by: Matthew Wilcox <willy@infradead.org>
Reviewed-by: Matthew Wilcox <willy@infradead.org>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Laurent Dufour <ldufour@linux.vnet.ibm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
