In order to pave the way for free page reporting in virtualized
environments we will need a way to get pages out of the free lists and
identify those pages after they have been returned. To accomplish this,
this patch adds the concept of a Reported Buddy, which is essentially
meant to just be the Uptodate flag used in conjunction with the Buddy page
type.
To prevent the reported pages from leaking outside of the buddy lists I
added a check to clear the PageReported bit in the del_page_from_free_list
function. As a result any reported page that is split, merged, or
allocated will have the flag cleared prior to the PageBuddy value being
cleared.
The process for reporting pages is fairly simple. Once we free a page
that meets the minimum order for page reporting we will schedule a worker
thread to start 2s or more in the future. That worker thread will begin
working from the lowest supported page reporting order up to MAX_ORDER - 1
pulling unreported pages from the free list and storing them in the
scatterlist.
When processing each individual free list it is necessary for the worker
thread to release the zone lock when it needs to stop and report the full
scatterlist of pages. To reduce the work of the next iteration the worker
thread will rotate the free list so that the first unreported page in the
free list becomes the first entry in the list.
It will then call a reporting function providing information on how many
entries are in the scatterlist. Once the function completes it will
return the pages to the free area from which they were allocated and start
over pulling more pages from the free areas until there are no longer
enough pages to report on to keep the worker busy, or we have processed as
many pages as were contained in the free area when we started processing
the list.
The worker thread will work in a round-robin fashion making its way though
each zone requesting reporting, and through each reportable free list
within that zone. Once all free areas within the zone have been processed
it will check to see if there have been any requests for reporting while
it was processing. If so it will reschedule the worker thread to start up
again in roughly 2s and exit.
Signed-off-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Nitesh Narayan Lal <nitesh@redhat.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Pankaj Gupta <pagupta@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Wei Wang <wei.w.wang@intel.com>
Cc: Yang Zhang <yang.zhang.wz@gmail.com>
Cc: wei qi <weiqi4@huawei.com>
Link: http://lkml.kernel.org/r/20200211224635.29318.19750.stgit@localhost.localdomain
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm / virtio: Provide support for free page reporting", v17.
This series provides an asynchronous means of reporting free guest pages
to a hypervisor so that the memory associated with those pages can be
dropped and reused by other processes and/or guests on the host. Using
this it is possible to avoid unnecessary I/O to disk and greatly improve
performance in the case of memory overcommit on the host.
When enabled we will be performing a scan of free memory every 2 seconds
while pages of sufficiently high order are being freed. In each pass at
least one sixteenth of each free list will be reported. By doing this we
avoid racing against other threads that may be causing a high amount of
memory churn.
The lowest page order currently scanned when reporting pages is
pageblock_order so that this feature will not interfere with the use of
Transparent Huge Pages in the case of virtualization.
Currently this is only in use by virtio-balloon however there is the hope
that at some point in the future other hypervisors might be able to make
use of it. In the virtio-balloon/QEMU implementation the hypervisor is
currently using MADV_DONTNEED to indicate to the host kernel that the page
is currently free. It will be zeroed and faulted back into the guest the
next time the page is accessed.
To track if a page is reported or not the Uptodate flag was repurposed and
used as a Reported flag for Buddy pages. We walk though the free list
isolating pages and adding them to the scatterlist until we either
encounter the end of the list or have processed at least one sixteenth of
the pages that were listed in nr_free prior to us starting. If we fill
the scatterlist before we reach the end of the list we rotate the list so
that the first unreported page we encounter is moved to the head of the
list as that is where we will resume after we have freed the reported
pages back into the tail of the list.
Below are the results from various benchmarks. I primarily focused on two
tests. The first is the will-it-scale/page_fault2 test, and the other is
a modified version of will-it-scale/page_fault1 that was enabled to use
THP. I did this as it allows for better visibility into different parts
of the memory subsystem. The guest is running with 32G for RAM on one
node of a E5-2630 v3. The host has had some features such as CPU turbo
disabled in the BIOS.
Test page_fault1 (THP) page_fault2
Name tasks Process Iter STDEV Process Iter STDEV
Baseline 1 1012402.50 0.14% 361855.25 0.81%
16 8827457.25 0.09% 3282347.00 0.34%
Patches Applied 1 1007897.00 0.23% 361887.00 0.26%
16 8784741.75 0.39% 3240669.25 0.48%
Patches Enabled 1 1010227.50 0.39% 359749.25 0.56%
16 8756219.00 0.24% 3226608.75 0.97%
Patches Enabled 1 1050982.00 4.26% 357966.25 0.14%
page shuffle 16 8672601.25 0.49% 3223177.75 0.40%
Patches enabled 1 1003238.00 0.22% 360211.00 0.22%
shuffle w/ RFC 16 8767010.50 0.32% 3199874.00 0.71%
The results above are for a baseline with a linux-next-20191219 kernel,
that kernel with this patch set applied but page reporting disabled in
virtio-balloon, the patches applied and page reporting fully enabled, the
patches enabled with page shuffling enabled, and the patches applied with
page shuffling enabled and an RFC patch that makes used of MADV_FREE in
QEMU. These results include the deviation seen between the average value
reported here versus the high and/or low value. I observed that during
the test memory usage for the first three tests never dropped whereas with
the patches fully enabled the VM would drop to using only a few GB of the
host's memory when switching from memhog to page fault tests.
Any of the overhead visible with this patch set enabled seems due to page
faults caused by accessing the reported pages and the host zeroing the
page before giving it back to the guest. This overhead is much more
visible when using THP than with standard 4K pages. In addition page
shuffling seemed to increase the amount of faults generated due to an
increase in memory churn. The overehad is reduced when using MADV_FREE as
we can avoid the extra zeroing of the pages when they are reintroduced to
the host, as can be seen when the RFC is applied with shuffling enabled.
The overall guest size is kept fairly small to only a few GB while the
test is running. If the host memory were oversubscribed this patch set
should result in a performance improvement as swapping memory in the host
can be avoided.
A brief history on the background of free page reporting can be found at:
https://lore.kernel.org/lkml/29f43d5796feed0dec8e8bb98b187d9dac03b900.camel@linux.intel.com/
This patch (of 9):
Move the head/tail adding logic out of the shuffle code and into the
__free_one_page function since ultimately that is where it is really
needed anyway. By doing this we should be able to reduce the overhead and
can consolidate all of the list addition bits in one spot.
Signed-off-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Yang Zhang <yang.zhang.wz@gmail.com>
Cc: Pankaj Gupta <pagupta@redhat.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Nitesh Narayan Lal <nitesh@redhat.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Wei Wang <wei.w.wang@intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: wei qi <weiqi4@huawei.com>
Link: http://lkml.kernel.org/r/20200211224602.29318.84523.stgit@localhost.localdomain
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "fix THP migration for CMA allocations", v2.
Transparent huge pages are allocated with __GFP_MOVABLE, and can end up in
CMA memory blocks. Transparent huge pages also have most of the
infrastructure in place to allow migration.
However, a few pieces were missing, causing THP migration to fail when
attempting to use CMA to allocate 1GB hugepages.
With these patches in place, THP migration from CMA blocks seems to work,
both for anonymous THPs and for tmpfs/shmem THPs.
This patch (of 2):
Add information to struct compact_control to indicate that the allocator
would really like to clear out this specific part of memory, used by for
example CMA.
Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Link: http://lkml.kernel.org/r/20200227213238.1298752-1-riel@surriel.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch makes ALLOC_KSWAPD equal to __GFP_KSWAPD_RECLAIM (cast to int).
Thanks to that code like:
if (gfp_mask & __GFP_KSWAPD_RECLAIM)
alloc_flags |= ALLOC_KSWAPD;
can be changed to:
alloc_flags |= (__force int) (gfp_mask &__GFP_KSWAPD_RECLAIM);
Thanks to this one branch less is generated in the assembly.
In case of ALLOC_KSWAPD flag two branches are saved, first one in code
that always executes in the beginning of page allocation and the second
one in loop in page allocator slowpath.
Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: http://lkml.kernel.org/r/20200304162118.14784-1-mateusznosek0@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently, the vm.min_free_kbytes sysctl value is capped at a hardcoded
64M in init_per_zone_wmark_min (unless it is overridden by khugepaged
initialization).
This value has not been modified since 2005, and enterprise-grade systems
now frequently have hundreds of GB of RAM and multiple 10, 40, or even 100
GB NICs. We have seen page allocation failures on heavily loaded systems
related to NIC drivers. These issues were resolved by an increase to
vm.min_free_kbytes.
This patch increases the hardcoded value by a factor of 4 as a temporary
solution.
Further work to make the calculation of vm.min_free_kbytes more consistent
throughout the kernel would be desirable.
As an example of the inconsistency of the current method, this value is
recalculated by init_per_zone_wmark_min() in the case of memory hotplug
which will override the value set by set_recommended_min_free_kbytes()
called during khugepaged initialization even if khugepaged remains
enabled, however an on/off toggle of khugepaged will then recalculate and
set the value via set_recommended_min_free_kbytes().
Signed-off-by: Joel Savitz <jsavitz@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Rafael Aquini <aquini@redhat.com>
Link: http://lkml.kernel.org/r/20200220150103.5183-1-jsavitz@redhat.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For huge pages (and in fact, any compound page), the GUP_PIN_COUNTING_BIAS
scheme tends to overflow too easily, each tail page increments the head
page->_refcount by GUP_PIN_COUNTING_BIAS (1024). That limits the number
of huge pages that can be pinned.
This patch removes that limitation, by using an exact form of pin counting
for compound pages of order > 1. The "order > 1" is required because this
approach uses the 3rd struct page in the compound page, and order 1
compound pages only have two pages, so that won't work there.
A new struct page field, hpage_pinned_refcount, has been added, replacing
a padding field in the union (so no new space is used).
This enhancement also has a useful side effect: huge pages and compound
pages (of order > 1) do not suffer from the "potential false positives"
problem that is discussed in the page_dma_pinned() comment block. That is
because these compound pages have extra space for tracking things, so they
get exact pin counts instead of overloading page->_refcount.
Documentation/core-api/pin_user_pages.rst is updated accordingly.
Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Link: http://lkml.kernel.org/r/20200211001536.1027652-8-jhubbard@nvidia.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Another "small" merge point to handle conflicts in a sane way.
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I5dc2f5f11275b29f3c9b5b8d4dd59864ceb6faf9
Patch series "mm/memory_hotplug: Shrink zones before removing memory", v6.
This series fixes the access of uninitialized memmaps when shrinking
zones/nodes and when removing memory. Also, it contains all fixes for
crashes that can be triggered when removing certain namespace using
memunmap_pages() - ZONE_DEVICE, reported by Aneesh.
We stop trying to shrink ZONE_DEVICE, as it's buggy, fixing it would be
more involved (we don't have SECTION_IS_ONLINE as an indicator), and
shrinking is only of limited use (set_zone_contiguous() cannot detect the
ZONE_DEVICE as contiguous).
We continue shrinking !ZONE_DEVICE zones, however, I reduced the amount of
code to a minimum. Shrinking is especially necessary to keep
zone->contiguous set where possible, especially, on memory unplug of DIMMs
at zone boundaries.
--------------------------------------------------------------------------
Zones are now properly shrunk when offlining memory blocks or when
onlining failed. This allows to properly shrink zones on memory unplug
even if the separate memory blocks of a DIMM were onlined to different
zones or re-onlined to a different zone after offlining.
Example:
:/# cat /proc/zoneinfo
Node 1, zone Movable
spanned 0
present 0
managed 0
:/# echo "online_movable" > /sys/devices/system/memory/memory41/state
:/# echo "online_movable" > /sys/devices/system/memory/memory43/state
:/# cat /proc/zoneinfo
Node 1, zone Movable
spanned 98304
present 65536
managed 65536
:/# echo 0 > /sys/devices/system/memory/memory43/online
:/# cat /proc/zoneinfo
Node 1, zone Movable
spanned 32768
present 32768
managed 32768
:/# echo 0 > /sys/devices/system/memory/memory41/online
:/# cat /proc/zoneinfo
Node 1, zone Movable
spanned 0
present 0
managed 0
This patch (of 6):
The third argument is actually number of pages. Change the variable name
from size to nr_pages to indicate this better.
No functional change in this patch.
Link: http://lkml.kernel.org/r/20191006085646.5768-3-david@redhat.com
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pankaj Gupta <pagupta@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.ibm.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Logan Gunthorpe <logang@deltatee.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 make sure that all memory holes are actually marked PageReserved(),
that page_to_pfn() produces reliable results, and that these pages are not
detected as "mmap" pages due to the mapcount.
E.g., booting a x86-64 QEMU guest with 4160 MB:
[ 0.010585] Early memory node ranges
[ 0.010586] node 0: [mem 0x0000000000001000-0x000000000009efff]
[ 0.010588] node 0: [mem 0x0000000000100000-0x00000000bffdefff]
[ 0.010589] node 0: [mem 0x0000000100000000-0x0000000143ffffff]
max_pfn is 0x144000.
Before this change:
[root@localhost ~]# ./page-types -r -a 0x144000,
flags page-count MB symbolic-flags long-symbolic-flags
0x0000000000000800 16384 64 ___________M_______________________________ mmap
total 16384 64
After this change:
[root@localhost ~]# ./page-types -r -a 0x144000,
flags page-count MB symbolic-flags long-symbolic-flags
0x0000000100000000 16384 64 ___________________________r_______________ reserved
total 16384 64
IOW, especially the unavailable physical memory ("memory hole") in the
last section would not get properly marked PageReserved() and is indicated
to be "mmap" memory.
Drop the trace of that function from include/linux/mm.h - nobody else
needs it, and rename it accordingly.
Note: The fake zone/node might not be covered by the zone/node span. This
is not an urgent issue (for now, we had the same node/zone due to the
zeroing). We'll need a clean way to mark memory holes (e.g., using a page
type PageHole() if possible or a fake ZONE_INVALID) and eventually stop
marking these memory holes PageReserved().
Link: http://lkml.kernel.org/r/20191211163201.17179-4-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Bob Picco <bob.picco@oracle.com>
Cc: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm: fix max_pfn not falling on section boundary", v2.
Playing with different memory sizes for a x86-64 guest, I discovered that
some memmaps (highest section if max_mem does not fall on the section
boundary) are marked as being valid and online, but contain garbage. We
have to properly initialize these memmaps.
Looking at /proc/kpageflags and friends, I found some more issues,
partially related to this.
This patch (of 3):
If max_pfn is not aligned to a section boundary, we can easily run into
BUGs. This can e.g., be triggered on x86-64 under QEMU by specifying a
memory size that is not a multiple of 128MB (e.g., 4097MB, but also
4160MB). I was told that on real HW, we can easily have this scenario
(esp., one of the main reasons sub-section hotadd of devmem was added).
The issue is, that we have a valid memmap (pfn_valid()) for the whole
section, and the whole section will be marked "online".
pfn_to_online_page() will succeed, but the memmap contains garbage.
E.g., doing a "./page-types -r -a 0x144001" when QEMU was started with "-m
4160M" - (see tools/vm/page-types.c):
[ 200.476376] BUG: unable to handle page fault for address: fffffffffffffffe
[ 200.477500] #PF: supervisor read access in kernel mode
[ 200.478334] #PF: error_code(0x0000) - not-present page
[ 200.479076] PGD 59614067 P4D 59614067 PUD 59616067 PMD 0
[ 200.479557] Oops: 0000 [#4] SMP NOPTI
[ 200.479875] CPU: 0 PID: 603 Comm: page-types Tainted: G D W 5.5.0-rc1-next-20191209 #93
[ 200.480646] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu4
[ 200.481648] RIP: 0010:stable_page_flags+0x4d/0x410
[ 200.482061] Code: f3 ff 41 89 c0 48 b8 00 00 00 00 01 00 00 00 45 84 c0 0f 85 cd 02 00 00 48 8b 53 08 48 8b 2b 48f
[ 200.483644] RSP: 0018:ffffb139401cbe60 EFLAGS: 00010202
[ 200.484091] RAX: fffffffffffffffe RBX: fffffbeec5100040 RCX: 0000000000000000
[ 200.484697] RDX: 0000000000000001 RSI: ffffffff9535c7cd RDI: 0000000000000246
[ 200.485313] RBP: ffffffffffffffff R08: 0000000000000000 R09: 0000000000000000
[ 200.485917] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000144001
[ 200.486523] R13: 00007ffd6ba55f48 R14: 00007ffd6ba55f40 R15: ffffb139401cbf08
[ 200.487130] FS: 00007f68df717580(0000) GS:ffff9ec77fa00000(0000) knlGS:0000000000000000
[ 200.487804] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 200.488295] CR2: fffffffffffffffe CR3: 0000000135d48000 CR4: 00000000000006f0
[ 200.488897] Call Trace:
[ 200.489115] kpageflags_read+0xe9/0x140
[ 200.489447] proc_reg_read+0x3c/0x60
[ 200.489755] vfs_read+0xc2/0x170
[ 200.490037] ksys_pread64+0x65/0xa0
[ 200.490352] do_syscall_64+0x5c/0xa0
[ 200.490665] entry_SYSCALL_64_after_hwframe+0x49/0xbe
But it can be triggered much easier via "cat /proc/kpageflags > /dev/null"
after cold/hot plugging a DIMM to such a system:
[root@localhost ~]# cat /proc/kpageflags > /dev/null
[ 111.517275] BUG: unable to handle page fault for address: fffffffffffffffe
[ 111.517907] #PF: supervisor read access in kernel mode
[ 111.518333] #PF: error_code(0x0000) - not-present page
[ 111.518771] PGD a240e067 P4D a240e067 PUD a2410067 PMD 0
This patch fixes that by at least zero-ing out that memmap (so e.g.,
page_to_pfn() will not crash). Commit 907ec5fca3 ("mm: zero remaining
unavailable struct pages") tried to fix a similar issue, but forgot to
consider this special case.
After this patch, there are still problems to solve. E.g., not all of
these pages falling into a memory hole will actually get initialized later
and set PageReserved - they are only zeroed out - but at least the
immediate crashes are gone. A follow-up patch will take care of this.
Link: http://lkml.kernel.org/r/20191211163201.17179-2-david@redhat.com
Fixes: f7f99100d8 ("mm: stop zeroing memory during allocation in vmemmap")
Signed-off-by: David Hildenbrand <david@redhat.com>
Tested-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Steven Sistare <steven.sistare@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Daniel Jordan <daniel.m.jordan@oracle.com>
Cc: Bob Picco <bob.picco@oracle.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: <stable@vger.kernel.org> [4.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It makes sense to call the WARN_ON_ONCE(zone_idx(zone) == ZONE_MOVABLE)
from start_isolate_page_range(), but should avoid triggering it from
userspace, i.e, from is_mem_section_removable() because it could crash
the system by a non-root user if warn_on_panic is set.
While at it, simplify the code a bit by removing an unnecessary jump
label.
Link: http://lkml.kernel.org/r/20200120163915.1469-1-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Suggested-by: Michal Hocko <mhocko@kernel.org>
Acked-by: Michal Hocko <mhocko@suse.com>
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>
It is not that hard to trigger lockdep splats by calling printk from
under zone->lock. Most of them are false positives caused by lock
chains introduced early in the boot process and they do not cause any
real problems (although most of the early boot lock dependencies could
happen after boot as well). There are some console drivers which do
allocate from the printk context as well and those should be fixed. In
any case, false positives are not that trivial to workaround and it is
far from optimal to lose lockdep functionality for something that is a
non-issue.
So change has_unmovable_pages() so that it no longer calls dump_page()
itself - instead it returns a "struct page *" of the unmovable page back
to the caller so that in the case of a has_unmovable_pages() failure,
the caller can call dump_page() after releasing zone->lock. Also, make
dump_page() is able to report a CMA page as well, so the reason string
from has_unmovable_pages() can be removed.
Even though has_unmovable_pages doesn't hold any reference to the
returned page this should be reasonably safe for the purpose of
reporting the page (dump_page) because it cannot be hotremoved in the
context of memory unplug. The state of the page might change but that
is the case even with the existing code as zone->lock only plays role
for free pages.
While at it, remove a similar but unnecessary debug-only printk() as
well. A sample of one of those lockdep splats is,
WARNING: possible circular locking dependency detected
------------------------------------------------------
test.sh/8653 is trying to acquire lock:
ffffffff865a4460 (console_owner){-.-.}, at:
console_unlock+0x207/0x750
but task is already holding lock:
ffff88883fff3c58 (&(&zone->lock)->rlock){-.-.}, at:
__offline_isolated_pages+0x179/0x3e0
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #3 (&(&zone->lock)->rlock){-.-.}:
__lock_acquire+0x5b3/0xb40
lock_acquire+0x126/0x280
_raw_spin_lock+0x2f/0x40
rmqueue_bulk.constprop.21+0xb6/0x1160
get_page_from_freelist+0x898/0x22c0
__alloc_pages_nodemask+0x2f3/0x1cd0
alloc_pages_current+0x9c/0x110
allocate_slab+0x4c6/0x19c0
new_slab+0x46/0x70
___slab_alloc+0x58b/0x960
__slab_alloc+0x43/0x70
__kmalloc+0x3ad/0x4b0
__tty_buffer_request_room+0x100/0x250
tty_insert_flip_string_fixed_flag+0x67/0x110
pty_write+0xa2/0xf0
n_tty_write+0x36b/0x7b0
tty_write+0x284/0x4c0
__vfs_write+0x50/0xa0
vfs_write+0x105/0x290
redirected_tty_write+0x6a/0xc0
do_iter_write+0x248/0x2a0
vfs_writev+0x106/0x1e0
do_writev+0xd4/0x180
__x64_sys_writev+0x45/0x50
do_syscall_64+0xcc/0x76c
entry_SYSCALL_64_after_hwframe+0x49/0xbe
-> #2 (&(&port->lock)->rlock){-.-.}:
__lock_acquire+0x5b3/0xb40
lock_acquire+0x126/0x280
_raw_spin_lock_irqsave+0x3a/0x50
tty_port_tty_get+0x20/0x60
tty_port_default_wakeup+0xf/0x30
tty_port_tty_wakeup+0x39/0x40
uart_write_wakeup+0x2a/0x40
serial8250_tx_chars+0x22e/0x440
serial8250_handle_irq.part.8+0x14a/0x170
serial8250_default_handle_irq+0x5c/0x90
serial8250_interrupt+0xa6/0x130
__handle_irq_event_percpu+0x78/0x4f0
handle_irq_event_percpu+0x70/0x100
handle_irq_event+0x5a/0x8b
handle_edge_irq+0x117/0x370
do_IRQ+0x9e/0x1e0
ret_from_intr+0x0/0x2a
cpuidle_enter_state+0x156/0x8e0
cpuidle_enter+0x41/0x70
call_cpuidle+0x5e/0x90
do_idle+0x333/0x370
cpu_startup_entry+0x1d/0x1f
start_secondary+0x290/0x330
secondary_startup_64+0xb6/0xc0
-> #1 (&port_lock_key){-.-.}:
__lock_acquire+0x5b3/0xb40
lock_acquire+0x126/0x280
_raw_spin_lock_irqsave+0x3a/0x50
serial8250_console_write+0x3e4/0x450
univ8250_console_write+0x4b/0x60
console_unlock+0x501/0x750
vprintk_emit+0x10d/0x340
vprintk_default+0x1f/0x30
vprintk_func+0x44/0xd4
printk+0x9f/0xc5
-> #0 (console_owner){-.-.}:
check_prev_add+0x107/0xea0
validate_chain+0x8fc/0x1200
__lock_acquire+0x5b3/0xb40
lock_acquire+0x126/0x280
console_unlock+0x269/0x750
vprintk_emit+0x10d/0x340
vprintk_default+0x1f/0x30
vprintk_func+0x44/0xd4
printk+0x9f/0xc5
__offline_isolated_pages.cold.52+0x2f/0x30a
offline_isolated_pages_cb+0x17/0x30
walk_system_ram_range+0xda/0x160
__offline_pages+0x79c/0xa10
offline_pages+0x11/0x20
memory_subsys_offline+0x7e/0xc0
device_offline+0xd5/0x110
state_store+0xc6/0xe0
dev_attr_store+0x3f/0x60
sysfs_kf_write+0x89/0xb0
kernfs_fop_write+0x188/0x240
__vfs_write+0x50/0xa0
vfs_write+0x105/0x290
ksys_write+0xc6/0x160
__x64_sys_write+0x43/0x50
do_syscall_64+0xcc/0x76c
entry_SYSCALL_64_after_hwframe+0x49/0xbe
other info that might help us debug this:
Chain exists of:
console_owner --> &(&port->lock)->rlock --> &(&zone->lock)->rlock
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&(&zone->lock)->rlock);
lock(&(&port->lock)->rlock);
lock(&(&zone->lock)->rlock);
lock(console_owner);
*** DEADLOCK ***
9 locks held by test.sh/8653:
#0: ffff88839ba7d408 (sb_writers#4){.+.+}, at:
vfs_write+0x25f/0x290
#1: ffff888277618880 (&of->mutex){+.+.}, at:
kernfs_fop_write+0x128/0x240
#2: ffff8898131fc218 (kn->count#115){.+.+}, at:
kernfs_fop_write+0x138/0x240
#3: ffffffff86962a80 (device_hotplug_lock){+.+.}, at:
lock_device_hotplug_sysfs+0x16/0x50
#4: ffff8884374f4990 (&dev->mutex){....}, at:
device_offline+0x70/0x110
#5: ffffffff86515250 (cpu_hotplug_lock.rw_sem){++++}, at:
__offline_pages+0xbf/0xa10
#6: ffffffff867405f0 (mem_hotplug_lock.rw_sem){++++}, at:
percpu_down_write+0x87/0x2f0
#7: ffff88883fff3c58 (&(&zone->lock)->rlock){-.-.}, at:
__offline_isolated_pages+0x179/0x3e0
#8: ffffffff865a4920 (console_lock){+.+.}, at:
vprintk_emit+0x100/0x340
stack backtrace:
Hardware name: HPE ProLiant DL560 Gen10/ProLiant DL560 Gen10,
BIOS U34 05/21/2019
Call Trace:
dump_stack+0x86/0xca
print_circular_bug.cold.31+0x243/0x26e
check_noncircular+0x29e/0x2e0
check_prev_add+0x107/0xea0
validate_chain+0x8fc/0x1200
__lock_acquire+0x5b3/0xb40
lock_acquire+0x126/0x280
console_unlock+0x269/0x750
vprintk_emit+0x10d/0x340
vprintk_default+0x1f/0x30
vprintk_func+0x44/0xd4
printk+0x9f/0xc5
__offline_isolated_pages.cold.52+0x2f/0x30a
offline_isolated_pages_cb+0x17/0x30
walk_system_ram_range+0xda/0x160
__offline_pages+0x79c/0xa10
offline_pages+0x11/0x20
memory_subsys_offline+0x7e/0xc0
device_offline+0xd5/0x110
state_store+0xc6/0xe0
dev_attr_store+0x3f/0x60
sysfs_kf_write+0x89/0xb0
kernfs_fop_write+0x188/0x240
__vfs_write+0x50/0xa0
vfs_write+0x105/0x290
ksys_write+0xc6/0x160
__x64_sys_write+0x43/0x50
do_syscall_64+0xcc/0x76c
entry_SYSCALL_64_after_hwframe+0x49/0xbe
Link: http://lkml.kernel.org/r/20200117181200.20299-1-cai@lca.pw
Signed-off-by: Qian Cai <cai@lca.pw>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.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>
Commit 96a2b03f28 ("mm, debug_pagelloc: use static keys to enable
debugging") has introduced a static key to reduce overhead when
debug_pagealloc is compiled in but not enabled. It relied on the
assumption that jump_label_init() is called before parse_early_param()
as in start_kernel(), so when the "debug_pagealloc=on" option is parsed,
it is safe to enable the static key.
However, it turns out multiple architectures call parse_early_param()
earlier from their setup_arch(). x86 also calls jump_label_init() even
earlier, so no issue was found while testing the commit, but same is not
true for e.g. ppc64 and s390 where the kernel would not boot with
debug_pagealloc=on as found by our QA.
To fix this without tricky changes to init code of multiple
architectures, this patch partially reverts the static key conversion
from 96a2b03f28. Init-time and non-fastpath calls (such as in arch
code) of debug_pagealloc_enabled() will again test a simple bool
variable. Fastpath mm code is converted to a new
debug_pagealloc_enabled_static() variant that relies on the static key,
which is enabled in a well-defined point in mm_init() where it's
guaranteed that jump_label_init() has been called, regardless of
architecture.
[sfr@canb.auug.org.au: export _debug_pagealloc_enabled_early]
Link: http://lkml.kernel.org/r/20200106164944.063ac07b@canb.auug.org.au
Link: http://lkml.kernel.org/r/20191219130612.23171-1-vbabka@suse.cz
Fixes: 96a2b03f28 ("mm, debug_pagelloc: use static keys to enable debugging")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Qian Cai <cai@lca.pw>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
THP page faults now attempt a __GFP_THISNODE allocation first, which
should only compact existing free memory, followed by another attempt
that can allocate from any node using reclaim/compaction effort
specified by global defrag setting and madvise.
This patch makes the following changes to the scheme:
- Before the patch, the first allocation relies on a check for
pageblock order and __GFP_IO to prevent excessive reclaim. This
however affects also the second attempt, which is not limited to
single node.
Instead of that, reuse the existing check for costly order
__GFP_NORETRY allocations, and make sure the first THP attempt uses
__GFP_NORETRY. As a side-effect, all costly order __GFP_NORETRY
allocations will bail out if compaction needs reclaim, while
previously they only bailed out when compaction was deferred due to
previous failures.
This should be still acceptable within the __GFP_NORETRY semantics.
- Before the patch, the second allocation attempt (on all nodes) was
passing __GFP_NORETRY. This is redundant as the check for pageblock
order (discussed above) was stronger. It's also contrary to
madvise(MADV_HUGEPAGE) which means some effort to allocate THP is
requested.
After this patch, the second attempt doesn't pass __GFP_THISNODE nor
__GFP_NORETRY.
To sum up, THP page faults now try the following attempts:
1. local node only THP allocation with no reclaim, just compaction.
2. for madvised VMA's or when synchronous compaction is enabled always - THP
allocation from any node with effort determined by global defrag setting
and VMA madvise
3. fallback to base pages on any node
Link: http://lkml.kernel.org/r/08a3f4dd-c3ce-0009-86c5-9ee51aba8557@suse.cz
Fixes: b39d0ee263 ("mm, page_alloc: avoid expensive reclaim when compaction may not succeed")
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There is a per-memcg lruvec and a NUMA node lruvec. Which one is being
used is somewhat confusing right now, and it's easy to make mistakes -
especially when it comes to global reclaim.
How it works: when memory cgroups are enabled, we always use the
root_mem_cgroup's per-node lruvecs. When memory cgroups are not compiled
in or disabled at runtime, we use pgdat->lruvec.
Document that in a comment.
Due to the way the reclaim code is generalized, all lookups use the
mem_cgroup_lruvec() helper function, and nobody should have to find the
right lruvec manually right now. But to avoid future mistakes, rename the
pgdat->lruvec member to pgdat->__lruvec and delete the convenience wrapper
that suggests it's a commonly accessed member.
While in this area, swap the mem_cgroup_lruvec() argument order. The name
suggests a memcg operation, yet it takes a pgdat first and a memcg second.
I have to double take every time I call this. Fix that.
Link: http://lkml.kernel.org/r/20191022144803.302233-3-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Cc: Roman Gushchin <guro@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "mm: Memory offlining + page isolation cleanups", v2.
This patch (of 2):
We call __offline_isolated_pages() from __offline_pages() after all
pages were isolated and are either free (PageBuddy()) or PageHWPoison.
Nothing can stop us from offlining memory at this point.
In __offline_isolated_pages() we first set all affected memory sections
offline (offline_mem_sections(pfn, end_pfn)), to mark the memmap as
invalid (pfn_to_online_page() will no longer succeed), and then walk
over all pages to pull the free pages from the free lists (to the
isolated free lists, to be precise).
Note that re-onlining a memory block will result in the whole memmap
getting reinitialized, overwriting any old state. We already poision
the memmap when offlining is complete to find any access to
stale/uninitialized memmaps.
So, setting the pages PageReserved() is not helpful. The memap is
marked offline and all pageblocks are isolated. As soon as offline, the
memmap is stale either way.
This looks like a leftover from ancient times where we initialized the
memmap when adding memory and not when onlining it (the pages were set
PageReserved so re-onling would work as expected).
Link: http://lkml.kernel.org/r/20191021172353.3056-2-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Mike Rapoport <rppt@linux.ibm.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Alexander Duyck <alexander.h.duyck@linux.intel.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Pavel Tatashin <pavel.tatashin@microsoft.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Cc: Pingfan Liu <kernelfans@gmail.com>
Cc: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
While investigating a bug related to higher atomic allocation failures,
we noticed the failure warnings positively drowning the console, and in
our case trigger lockup warnings because of a serial console too slow to
handle all that output.
But even if we had a faster console, it's unclear what additional
information the current level of repetition provides.
Allocation failures happen for three reasons: The machine is OOM, the VM
is failing to handle reasonable requests, or somebody is making
unreasonable requests (and didn't acknowledge their opportunism with
__GFP_NOWARN). Having the memory dump, a callstack, and the ratelimit
stats on skipped failure warnings should provide enough information to
let users/admins/developers know whether something is wrong and point
them in the right direction for debugging, bpftracing etc.
Limit allocation failure warnings to one spew every ten seconds.
Link: http://lkml.kernel.org/r/20191028194906.26899-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
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>
Deferred memory initialisation updates zone->managed_pages during the
initialisation phase but before that finishes, the per-cpu page
allocator (pcpu) calculates the number of pages allocated/freed in
batches as well as the maximum number of pages allowed on a per-cpu
list. As zone->managed_pages is not up to date yet, the pcpu
initialisation calculates inappropriately low batch and high values.
This increases zone lock contention quite severely in some cases with
the degree of severity depending on how many CPUs share a local zone and
the size of the zone. A private report indicated that kernel build
times were excessive with extremely high system CPU usage. A perf
profile indicated that a large chunk of time was lost on zone->lock
contention.
This patch recalculates the pcpu batch and high values after deferred
initialisation completes for every populated zone in the system. It was
tested on a 2-socket AMD EPYC 2 machine using a kernel compilation
workload -- allmodconfig and all available CPUs.
mmtests configuration: config-workload-kernbench-max Configuration was
modified to build on a fresh XFS partition.
kernbench
5.4.0-rc3 5.4.0-rc3
vanilla resetpcpu-v2
Amean user-256 13249.50 ( 0.00%) 16401.31 * -23.79%*
Amean syst-256 14760.30 ( 0.00%) 4448.39 * 69.86%*
Amean elsp-256 162.42 ( 0.00%) 119.13 * 26.65%*
Stddev user-256 42.97 ( 0.00%) 19.15 ( 55.43%)
Stddev syst-256 336.87 ( 0.00%) 6.71 ( 98.01%)
Stddev elsp-256 2.46 ( 0.00%) 0.39 ( 84.03%)
5.4.0-rc3 5.4.0-rc3
vanilla resetpcpu-v2
Duration User 39766.24 49221.79
Duration System 44298.10 13361.67
Duration Elapsed 519.11 388.87
The patch reduces system CPU usage by 69.86% and total build time by
26.65%. The variance of system CPU usage is also much reduced.
Before, this was the breakdown of batch and high values over all zones
was:
256 batch: 1
256 batch: 63
512 batch: 7
256 high: 0
256 high: 378
512 high: 42
512 pcpu pagesets had a batch limit of 7 and a high limit of 42. After
the patch:
256 batch: 1
768 batch: 63
256 high: 0
768 high: 378
[mgorman@techsingularity.net: fix merge/linkage snafu]
Link: http://lkml.kernel.org/r/20191023084705.GD3016@techsingularity.netLink: http://lkml.kernel.org/r/20191021094808.28824-2-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Matt Fleming <matt@codeblueprint.co.uk>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Qian Cai <cai@lca.pw>
Cc: <stable@vger.kernel.org> [4.1+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit b39d0ee263 ("mm, page_alloc: avoid expensive reclaim when
compaction may not succeed") has chnaged the allocator to bail out from
the allocator early to prevent from a potentially excessive memory
reclaim. __GFP_RETRY_MAYFAIL is designed to retry the allocation,
reclaim and compaction loop as long as there is a reasonable chance to
make forward progress. Neither COMPACT_SKIPPED nor COMPACT_DEFERRED at
the INIT_COMPACT_PRIORITY compaction attempt gives this feedback.
The most obvious affected subsystem is hugetlbfs which allocates huge
pages based on an admin request (or via admin configured overcommit). I
have done a simple test which tries to allocate half of the memory for
hugetlb pages while the memory is full of a clean page cache. This is
not an unusual situation because we try to cache as much of the memory
as possible and sysctl/sysfs interface to allocate huge pages is there
for flexibility to allocate hugetlb pages at any time.
System has 1GB of RAM and we are requesting 515MB worth of hugetlb pages
after the memory is prefilled by a clean page cache:
root@test1:~# cat hugetlb_test.sh
set -x
echo 0 > /proc/sys/vm/nr_hugepages
echo 3 > /proc/sys/vm/drop_caches
echo 1 > /proc/sys/vm/compact_memory
dd if=/mnt/data/file-1G of=/dev/null bs=$((4<<10))
TS=$(date +%s)
echo 256 > /proc/sys/vm/nr_hugepages
cat /proc/sys/vm/nr_hugepages
The results for 2 consecutive runs on clean 5.3
root@test1:~# sh hugetlb_test.sh
+ echo 0
+ echo 3
+ echo 1
+ dd if=/mnt/data/file-1G of=/dev/null bs=4096
262144+0 records in
262144+0 records out
1073741824 bytes (1.1 GB) copied, 21.0694 s, 51.0 MB/s
+ date +%s
+ TS=1569905284
+ echo 256
+ cat /proc/sys/vm/nr_hugepages
256
root@test1:~# sh hugetlb_test.sh
+ echo 0
+ echo 3
+ echo 1
+ dd if=/mnt/data/file-1G of=/dev/null bs=4096
262144+0 records in
262144+0 records out
1073741824 bytes (1.1 GB) copied, 21.7548 s, 49.4 MB/s
+ date +%s
+ TS=1569905311
+ echo 256
+ cat /proc/sys/vm/nr_hugepages
256
Now with b39d0ee263 applied
root@test1:~# sh hugetlb_test.sh
+ echo 0
+ echo 3
+ echo 1
+ dd if=/mnt/data/file-1G of=/dev/null bs=4096
262144+0 records in
262144+0 records out
1073741824 bytes (1.1 GB) copied, 20.1815 s, 53.2 MB/s
+ date +%s
+ TS=1569905516
+ echo 256
+ cat /proc/sys/vm/nr_hugepages
11
root@test1:~# sh hugetlb_test.sh
+ echo 0
+ echo 3
+ echo 1
+ dd if=/mnt/data/file-1G of=/dev/null bs=4096
262144+0 records in
262144+0 records out
1073741824 bytes (1.1 GB) copied, 21.9485 s, 48.9 MB/s
+ date +%s
+ TS=1569905541
+ echo 256
+ cat /proc/sys/vm/nr_hugepages
12
The success rate went down by factor of 20!
Although hugetlb allocation requests might fail and it is reasonable to
expect them to under extremely fragmented memory or when the memory is
under a heavy pressure but the above situation is not that case.
Fix the regression by reverting back to the previous behavior for
__GFP_RETRY_MAYFAIL requests and disable the beail out heuristic for
those requests.
Mike said:
: hugetlbfs allocations are commonly done via sysctl/sysfs shortly after
: boot where this may not be as much of an issue. However, I am aware of at
: least three use cases where allocations are made after the system has been
: up and running for quite some time:
:
: - DB reconfiguration. If sysctl/sysfs fails to get required number of
: huge pages, system is rebooted to perform allocation after boot.
:
: - VM provisioning. If unable get required number of huge pages, fall
: back to base pages.
:
: - An application that does not preallocate pool, but rather allocates
: pages at fault time for optimal NUMA locality.
:
: In all cases, I would expect b39d0ee263 to cause regressions and
: noticable behavior changes.
:
: My quick/limited testing in
: https://lkml.kernel.org/r/3468b605-a3a9-6978-9699-57c52a90bd7e@oracle.com
: was insufficient. It was also mentioned that if something like
: b39d0ee263 went forward, I would like exemptions for __GFP_RETRY_MAYFAIL
: requests as in this patch.
[mhocko@suse.com: reworded changelog]
Link: http://lkml.kernel.org/r/20191007075548.12456-1-mhocko@kernel.org
Fixes: b39d0ee263 ("mm, page_alloc: avoid expensive reclaim when compaction may not succeed")
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
To make the 5.4-rc1 merge easier, merge at a prerelease point in time
before the final release happens.
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: If613d657fd0abf9910c5bf3435a745f01b89765e
Merge hugepage allocation updates from David Rientjes:
"We (mostly Linus, Andrea, and myself) have been discussing offlist how
to implement a sane default allocation strategy for hugepages on NUMA
platforms.
With these reverts in place, the page allocator will happily allocate
a remote hugepage immediately rather than try to make a local hugepage
available. This incurs a substantial performance degradation when
memory compaction would have otherwise made a local hugepage
available.
This series reverts those reverts and attempts to propose a more sane
default allocation strategy specifically for hugepages. Andrea
acknowledges this is likely to fix the swap storms that he originally
reported that resulted in the patches that removed __GFP_THISNODE from
hugepage allocations.
The immediate goal is to return 5.3 to the behavior the kernel has
implemented over the past several years so that remote hugepages are
not immediately allocated when local hugepages could have been made
available because the increased access latency is untenable.
The next goal is to introduce a sane default allocation strategy for
hugepages allocations in general regardless of the configuration of
the system so that we prevent thrashing of local memory when
compaction is unlikely to succeed and can prefer remote hugepages over
remote native pages when the local node is low on memory."
Note on timing: this reverts the hugepage VM behavior changes that got
introduced fairly late in the 5.3 cycle, and that fixed a huge
performance regression for certain loads that had been around since
4.18.
Andrea had this note:
"The regression of 4.18 was that it was taking hours to start a VM
where 3.10 was only taking a few seconds, I reported all the details
on lkml when it was finally tracked down in August 2018.
https://lore.kernel.org/linux-mm/20180820032640.9896-2-aarcange@redhat.com/
__GFP_THISNODE in MADV_HUGEPAGE made the above enterprise vfio
workload degrade like in the "current upstream" above. And it still
would have been that bad as above until 5.3-rc5"
where the bad behavior ends up happening as you fill up a local node,
and without that change, you'd get into the nasty swap storm behavior
due to compaction working overtime to make room for more memory on the
nodes.
As a result 5.3 got the two performance fix reverts in rc5.
However, David Rientjes then noted that those performance fixes in turn
regressed performance for other loads - although not quite to the same
degree. He suggested reverting the reverts and instead replacing them
with two small changes to how hugepage allocations are done (patch
descriptions rephrased by me):
- "avoid expensive reclaim when compaction may not succeed": just admit
that the allocation failed when you're trying to allocate a huge-page
and compaction wasn't successful.
- "allow hugepage fallback to remote nodes when madvised": when that
node-local huge-page allocation failed, retry without forcing the
local node.
but by then I judged it too late to replace the fixes for a 5.3 release.
So 5.3 was released with behavior that harked back to the pre-4.18 logic.
But now we're in the merge window for 5.4, and we can see if this
alternate model fixes not just the horrendous swap storm behavior, but
also restores the performance regression that the late reverts caused.
Fingers crossed.
* emailed patches from David Rientjes <rientjes@google.com>:
mm, page_alloc: allow hugepage fallback to remote nodes when madvised
mm, page_alloc: avoid expensive reclaim when compaction may not succeed
Revert "Revert "Revert "mm, thp: consolidate THP gfp handling into alloc_hugepage_direct_gfpmask""
Revert "Revert "mm, thp: restore node-local hugepage allocations""
Memory compaction has a couple significant drawbacks as the allocation
order increases, specifically:
- isolate_freepages() is responsible for finding free pages to use as
migration targets and is implemented as a linear scan of memory
starting at the end of a zone,
- failing order-0 watermark checks in memory compaction does not account
for how far below the watermarks the zone actually is: to enable
migration, there must be *some* free memory available. Per the above,
watermarks are not always suffficient if isolate_freepages() cannot
find the free memory but it could require hundreds of MBs of reclaim to
even reach this threshold (read: potentially very expensive reclaim with
no indication compaction can be successful), and
- if compaction at this order has failed recently so that it does not even
run as a result of deferred compaction, looping through reclaim can often
be pointless.
For hugepage allocations, these are quite substantial drawbacks because
these are very high order allocations (order-9 on x86) and falling back to
doing reclaim can potentially be *very* expensive without any indication
that compaction would even be successful.
Reclaim itself is unlikely to free entire pageblocks and certainly no
reliance should be put on it to do so in isolation (recall lumpy reclaim).
This means we should avoid reclaim and simply fail hugepage allocation if
compaction is deferred.
It is also not helpful to thrash a zone by doing excessive reclaim if
compaction may not be able to access that memory. If order-0 watermarks
fail and the allocation order is sufficiently large, it is likely better
to fail the allocation rather than thrashing the zone.
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Stefan Priebe - Profihost AG <s.priebe@profihost.ag>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
