We keep track of several kernel memory stats (total kernel memory, page
tables, stack, vmalloc, etc) on multiple levels (global, per-node,
per-memcg, etc). These stats give insights to users to how much memory
is used by the kernel and for what purposes.
Currently, memory used by KVM mmu is not accounted in any of those
kernel memory stats. This patch series accounts the memory pages
used by KVM for page tables in those stats in a new
NR_SECONDARY_PAGETABLE stat. This stat can be later extended to account
for other types of secondary pages tables (e.g. iommu page tables).
KVM has a decent number of large allocations that aren't for page
tables, but for most of them, the number/size of those allocations
scales linearly with either the number of vCPUs or the amount of memory
assigned to the VM. KVM's secondary page table allocations do not scale
linearly, especially when nested virtualization is in use.
From a KVM perspective, NR_SECONDARY_PAGETABLE will scale with KVM's
per-VM pages_{4k,2m,1g} stats unless the guest is doing something
bizarre (e.g. accessing only 4kb chunks of 2mb pages so that KVM is
forced to allocate a large number of page tables even though the guest
isn't accessing that much memory). However, someone would need to either
understand how KVM works to make that connection, or know (or be told) to
go look at KVM's stats if they're running VMs to better decipher the stats.
Furthermore, having NR_PAGETABLE side-by-side with NR_SECONDARY_PAGETABLE
is informative. For example, when backing a VM with THP vs. HugeTLB,
NR_SECONDARY_PAGETABLE is roughly the same, but NR_PAGETABLE is an order
of magnitude higher with THP. So having this stat will at the very least
prove to be useful for understanding tradeoffs between VM backing types,
and likely even steer folks towards potential optimizations.
The original discussion with more details about the rationale:
https://lore.kernel.org/all/87ilqoi77b.wl-maz@kernel.org
This stat will be used by subsequent patches to count KVM mmu
memory usage.
Signed-off-by: Yosry Ahmed <yosryahmed@google.com>
Acked-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220823004639.2387269-2-yosryahmed@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Bug: 222044477
(cherry picked from commit ebc97a52b5d6cd5fb0c15a3fc9cdd6eb924646a1)
[vdonnefort@: Fix trivial documentation conflict]
Change-Id: I16976e21d2e68ebbcd49e9f1275055e81ec82881
Signed-off-by: Vincent Donnefort <vdonnefort@google.com>
Introduce vma_can_speculate(), which allows speculative handling for
VMAs mapping supported file types.
From do_handle_mm_fault(), speculative handling will follow through
__handle_mm_fault(), handle_pte_fault() and do_fault().
At this point, we expect speculative faults to continue through one of:
- do_read_fault(), fully implemented;
- do_cow_fault(), which might abort if missing anon vmas,
- do_shared_fault(), not implemented yet
(would require ->page_mkwrite() changes).
vma_can_speculate() provides an early abort for the do_shared_fault() case,
limiting the time spent on trying that unimplemented case.
Signed-off-by: Michel Lespinasse <michel@lespinasse.org>
Link: https://lore.kernel.org/all/20210407014502.24091-31-michel@lespinasse.org/
Conflicts:
include/linux/vm_event_item.h
mm/vmstat.c
1. SPF_ATTEMPT_FILE is taken from https://lore.kernel.org/all/20210407014502.24091-36-michel@lespinasse.org/
since the patch posted upstream at the time had a different structure
with stats for anonymouse and file-backed pagefaults introduced in a
separate patch.
Bug: 161210518
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Change-Id: I3a28af63b41b649f02f8b73d53f6494ad114ee5a
Attempt speculative mm fault handling first, and fall back to the
existing (non-speculative) code if that fails.
The speculative handling closely mirrors the non-speculative logic.
This includes some x86 specific bits such as the access_error() call.
This is why we chose to implement the speculative handling in arch/x86
rather than in common code.
The vma is first looked up and copied, under protection of the rcu
read lock. The mmap lock sequence count is used to verify the
integrity of the copied vma, and passed to do_handle_mm_fault() to
allow checking against races with mmap writers when finalizing the fault.
Signed-off-by: Michel Lespinasse <michel@lespinasse.org>
Link: https://lore.kernel.org/all/20220128131006.67712-14-michel@lespinasse.org/
Bug: 161210518
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Change-Id: I2c078a173ee39f35af16daeee8c6a1466d10c3e8
Disable preemption on -RT for the vmstat code. On vanila the code runs in
IRQ-off regions while on -RT it may not when stats are updated under a
local_lock. "preempt_disable" ensures that the same resources is not
updated in parallel due to preemption.
This patch differs from the preempt-rt version where __count_vm_event and
__count_vm_events are also protected. The counters are explicitly
"allowed to be to be racy" so there is no need to protect them from
preemption. Only the accurate page stats that are updated by a
read-modify-write need protection. This patch also differs in that a
preempt_[en|dis]able_rt helper is not used. As vmstat is the only user of
the helper, it was suggested that it be open-coded in vmstat.c instead of
risking the helper being used in unnecessary contexts.
Link: https://lkml.kernel.org/r/20210805160019.1137-2-mgorman@techsingularity.net
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merge misc updates from Andrew Morton:
"173 patches.
Subsystems affected by this series: ia64, ocfs2, block, and mm (debug,
pagecache, gup, swap, shmem, memcg, selftests, pagemap, mremap,
bootmem, sparsemem, vmalloc, kasan, pagealloc, memory-failure,
hugetlb, userfaultfd, vmscan, compaction, mempolicy, memblock,
oom-kill, migration, ksm, percpu, vmstat, and madvise)"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (173 commits)
mm/madvise: add MADV_WILLNEED to process_madvise()
mm/vmstat: remove unneeded return value
mm/vmstat: simplify the array size calculation
mm/vmstat: correct some wrong comments
mm/percpu,c: remove obsolete comments of pcpu_chunk_populated()
selftests: vm: add COW time test for KSM pages
selftests: vm: add KSM merging time test
mm: KSM: fix data type
selftests: vm: add KSM merging across nodes test
selftests: vm: add KSM zero page merging test
selftests: vm: add KSM unmerge test
selftests: vm: add KSM merge test
mm/migrate: correct kernel-doc notation
mm: wire up syscall process_mrelease
mm: introduce process_mrelease system call
memblock: make memblock_find_in_range method private
mm/mempolicy.c: use in_task() in mempolicy_slab_node()
mm/mempolicy: unify the create() func for bind/interleave/prefer-many policies
mm/mempolicy: advertise new MPOL_PREFERRED_MANY
mm/hugetlb: add support for mempolicy MPOL_PREFERRED_MANY
...
The functions get_online_cpus() and put_online_cpus() have been
deprecated during the CPU hotplug rework. They map directly to
cpus_read_lock() and cpus_read_unlock().
Replace deprecated CPU-hotplug functions with the official version.
The behavior remains unchanged.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20210803141621.780504-21-bigeasy@linutronix.de
NUMA statistics are maintained on the zone level for hits, misses, foreign
etc but nothing relies on them being perfectly accurate for functional
correctness. The counters are used by userspace to get a general overview
of a workloads NUMA behaviour but the page allocator incurs a high cost to
maintain perfect accuracy similar to what is required for a vmstat like
NR_FREE_PAGES. There even is a sysctl vm.numa_stat to allow userspace to
turn off the collection of NUMA statistics like NUMA_HIT.
This patch converts NUMA_HIT and friends to be NUMA events with similar
accuracy to VM events. There is a possibility that slight errors will be
introduced but the overall trend as seen by userspace will be similar.
The counters are no longer updated from vmstat_refresh context as it is
unnecessary overhead for counters that may never be read by userspace.
Note that counters could be maintained at the node level to save space but
it would have a user-visible impact due to /proc/zoneinfo.
[lkp@intel.com: Fix misplaced closing brace for !CONFIG_NUMA]
Link: https://lkml.kernel.org/r/20210512095458.30632-4-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jesper Dangaard Brouer <brouer@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The PCP (per-cpu page allocator in page_alloc.c) shares locking
requirements with vmstat and the zone lock which is inconvenient and
causes some issues. For example, the PCP list and vmstat share the same
per-cpu space meaning that it's possible that vmstat updates dirty cache
lines holding per-cpu lists across CPUs unless padding is used. Second,
PREEMPT_RT does not want to disable IRQs for too long in the page
allocator.
This series splits the locking requirements and uses locks types more
suitable for PREEMPT_RT, reduces the time when special locking is required
for stats and reduces the time when IRQs need to be disabled on
!PREEMPT_RT kernels.
Why local_lock? PREEMPT_RT considers the following sequence to be unsafe
as documented in Documentation/locking/locktypes.rst
local_irq_disable();
spin_lock(&lock);
The pcp allocator has this sequence for rmqueue_pcplist (local_irq_save)
-> __rmqueue_pcplist -> rmqueue_bulk (spin_lock). While it's possible to
separate this out, it generally means there are points where we enable
IRQs and reenable them again immediately. To prevent a migration and the
per-cpu pointer going stale, migrate_disable is also needed. That is a
custom lock that is similar, but worse, than local_lock. Furthermore, on
PREEMPT_RT, it's undesirable to leave IRQs disabled for too long. By
converting to local_lock which disables migration on PREEMPT_RT, the
locking requirements can be separated and start moving the protections for
PCP, stats and the zone lock to PREEMPT_RT-safe equivalent locking. As a
bonus, local_lock also means that PROVE_LOCKING does something useful.
After that, it's obvious that zone_statistics incurs too much overhead and
leaves IRQs disabled for longer than necessary on !PREEMPT_RT kernels.
zone_statistics uses perfectly accurate counters requiring IRQs be
disabled for parallel RMW sequences when inaccurate ones like vm_events
would do. The series makes the NUMA statistics (NUMA_HIT and friends)
inaccurate counters that then require no special protection on
!PREEMPT_RT.
The bulk page allocator can then do stat updates in bulk with IRQs enabled
which should improve the efficiency. Technically, this could have been
done without the local_lock and vmstat conversion work and the order
simply reflects the timing of when different series were implemented.
Finally, there are places where we conflate IRQs being disabled for the
PCP with the IRQ-safe zone spinlock. The remainder of the series reduces
the scope of what is protected by disabled IRQs on !PREEMPT_RT kernels.
By the end of the series, page_alloc.c does not call local_irq_save so the
locking scope is a bit clearer. The one exception is that modifying
NR_FREE_PAGES still happens in places where it's known the IRQs are
disabled as it's harmless for PREEMPT_RT and would be expensive to split
the locking there.
No performance data is included because despite the overhead of the stats,
it's within the noise for most workloads on !PREEMPT_RT. However, Jesper
Dangaard Brouer ran a page allocation microbenchmark on a E5-1650 v4 @
3.60GHz CPU on the first version of this series. Focusing on the array
variant of the bulk page allocator reveals the following.
(CPU: Intel(R) Xeon(R) CPU E5-1650 v4 @ 3.60GHz)
ARRAY variant: time_bulk_page_alloc_free_array: step=bulk size
Baseline Patched
1 56.383 54.225 (+3.83%)
2 40.047 35.492 (+11.38%)
3 37.339 32.643 (+12.58%)
4 35.578 30.992 (+12.89%)
8 33.592 29.606 (+11.87%)
16 32.362 28.532 (+11.85%)
32 31.476 27.728 (+11.91%)
64 30.633 27.252 (+11.04%)
128 30.596 27.090 (+11.46%)
While this is a positive outcome, the series is more likely to be
interesting to the RT people in terms of getting parts of the PREEMPT_RT
tree into mainline.
This patch (of 9):
The per-cpu page allocator lists and the per-cpu vmstat deltas are stored
in the same struct per_cpu_pages even though vmstats have no direct impact
on the per-cpu page lists. This is inconsistent because the vmstats for a
node are stored on a dedicated structure. The bigger issue is that the
per_cpu_pages structure is not cache-aligned and stat updates either cache
conflict with adjacent per-cpu lists incurring a runtime cost or padding
is required incurring a memory cost.
This patch splits the per-cpu pagelists and the vmstat deltas into
separate structures. It's mostly a mechanical conversion but some
variable renaming is done to clearly distinguish the per-cpu pages
structure (pcp) from the vmstats (pzstats).
Superficially, this appears to increase the size of the per_cpu_pages
structure but the movement of expire fills a structure hole so there is no
impact overall.
[mgorman@techsingularity.net: make it W=1 cleaner]
Link: https://lkml.kernel.org/r/20210514144622.GA3735@techsingularity.net
[mgorman@techsingularity.net: make it W=1 even cleaner]
Link: https://lkml.kernel.org/r/20210516140705.GB3735@techsingularity.net
[lkp@intel.com: check struct per_cpu_zonestat has a non-zero size]
[vbabka@suse.cz: Init zone->per_cpu_zonestats properly]
Link: https://lkml.kernel.org/r/20210512095458.30632-1-mgorman@techsingularity.net
Link: https://lkml.kernel.org/r/20210512095458.30632-2-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Chuck Lever <chuck.lever@oracle.com>
Cc: Jesper Dangaard Brouer <brouer@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
One giant leap, all the way up to 5.13-rc1
Also take the opportunity to re-align (a.k.a. fix a couple of previous
merge conflict fix-up issues) which occurred during this merge-window.
Fixes: 4797acfb9c ("Merge 16b3d0cf5b Merge tag 'sched-core-2021-04-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip into android-mainline")
Fixes: 92f282f338 ("Merge 8ca5297e7e Merge tag 'kconfig-v5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild into android-mainline")
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Change-Id: Ie9389f595776e8f66bba6eaf0fa7a3587c6a5749
To help with debugging the sluggishness caused by TLB miss/reload, we
introduce monotonic hugepage [direct mapped] split event counts since
system state: SYSTEM_RUNNING to be displayed as part of /proc/vmstat in
x86 servers
The lifetime split event information will be displayed at the bottom of
/proc/vmstat
....
swap_ra 0
swap_ra_hit 0
direct_map_level2_splits 94
direct_map_level3_splits 4
nr_unstable 0
....
One of the many lasting sources of direct hugepage splits is kernel
tracing (kprobes, tracepoints).
Note that the kernel's code segment [512 MB] points to the same physical
addresses that have been already mapped in the kernel's direct mapping
range.
Source : Documentation/x86/x86_64/mm.rst
When we enable kernel tracing, the kernel has to modify
attributes/permissions of the text segment hugepages that are direct
mapped causing them to split.
Kernel's direct mapped hugepages do not coalesce back after split and
remain in place for the remainder of the lifetime.
An instance of direct page splits when we turn on dynamic kernel tracing
....
cat /proc/vmstat | grep -i direct_map_level
direct_map_level2_splits 784
direct_map_level3_splits 12
bpftrace -e 'tracepoint:raw_syscalls:sys_enter { @ [pid, comm] =
count(); }'
cat /proc/vmstat | grep -i
direct_map_level
direct_map_level2_splits 789
direct_map_level3_splits 12
....
Link: https://lkml.kernel.org/r/20210218235744.1040634-1-saravanand@fb.com
Signed-off-by: Saravanan D <saravanand@fb.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
vmstat_refresh() can occasionally catch nr_zone_write_pending and
nr_writeback when they are transiently negative. The reason is partly
that the interrupt which decrements them in test_clear_page_writeback()
can come in before __test_set_page_writeback() got to increment them;
but transient negatives are still seen even when that is prevented, and
I am not yet certain why (but see Roman's note below). Those stats are
not buggy, they have never been seen to drift away from 0 permanently:
so just avoid the annoyance of showing a warning on them.
Similarly avoid showing a warning on nr_free_cma: CMA users have seen
that one reported negative from /proc/sys/vm/stat_refresh too, but it
does drift away permanently: I believe that's because its incrementation
and decrementation are decided by page migratetype, but the migratetype
of a pageblock is not guaranteed to be constant.
Roman Gushchin points out:
"For performance reasons, vmstat counters are incremented and
decremented using per-cpu batches. vmstat_refresh() flushes the
per-cpu batches on all CPUs, to get values as accurate as possible;
but this method is not atomic, so the resulting value is not always
precise.
As a consequence, for those counters whose actual value is close to 0,
a small negative value may occasionally be reported. If the value is
small and the state is transient, it is not an indication of an error"
Link: https://lore.kernel.org/linux-mm/20200714173747.3315771-1-guro@fb.com/
Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2103012158540.7549@eggly.anvils
Signed-off-by: Hugh Dickins <hughd@google.com>
Reported-by: Roman Gushchin <guro@fb.com>
Acked-by: Roman Gushchin <guro@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Steps on the way to 5.12-rc1
Resolves merge conflicts in:
drivers/dma-buf/dma-heap.c
drivers/dma-buf/heaps/cma_heap.c
drivers/dma-buf/heaps/system_heap.c
include/linux/dma-heap.h
Cc: John Stultz <john.stultz@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ibb32dbdba5183c9e19f5d1e94016cc1ae9616173
Byte-accounted items are used for slab object accounting at the cgroup
level, because the objects in a slab page can belong to different cgroups.
At the global level these items always change in multiples of whole slab
pages. The vmstat code exploits this and stores these items as pages
internally, which allows for more compact per-cpu data.
This optimization isn't self-evident from the asserts and the division in
the stat update functions. Provide the reader with some context.
Link: https://lkml.kernel.org/r/20210202184411.118614-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On NOHZ, the periodic vmstat flushers on each CPU can go to sleep and
won't wake up until stat changes are detected in the per-cpu deltas of the
zone vmstat counters.
In commit 75ef718405 ("mm, vmstat: add infrastructure for per-node
vmstats") per-node counters were introduced, and subsequently most stats
were moved from the zone to the node level. However, the node counters
weren't added to the NOHZ wakeup detection.
In theory this can cause per-cpu errors to remain in the user-reported
stats indefinitely. In practice this only affects a handful of sub
counters (file_mapped, dirty and writeback e.g.) because other page state
changes at the node level likely involve a change at the zone level as
well (alloc and free, lru ops). Also, nobody has complained.
Fix it up for completeness: wake up vmstat refreshing on node changes.
Also remove the BUILD_BUG_ONs that assert counter size; we haven't relied
on it since we added sizeof() to the range calculation in commit
13c9aaf7fa ("mm/vmstat.c: fix NUMA statistics updates").
Link: https://lkml.kernel.org/r/20210202184342.118513-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Let's count the number of CMA pages per zone and print them in
/proc/zoneinfo.
Having access to the total number of CMA pages per zone is helpful for
debugging purposes to know where exactly the CMA pages ended up, and to
figure out how many pages of a zone might behave differently, even after
some of these pages might already have been allocated.
As one example, CMA pages part of a kernel zone cannot be used for
ordinary kernel allocations but instead behave more like ZONE_MOVABLE.
For now, we are only able to get the global nr+free cma pages from
/proc/meminfo and the free cma pages per zone from /proc/zoneinfo.
Example after this patch when booting a 6 GiB QEMU VM with
"hugetlb_cma=2G":
# cat /proc/zoneinfo | grep cma
cma 0
nr_free_cma 0
cma 0
nr_free_cma 0
cma 524288
nr_free_cma 493016
cma 0
cma 0
# cat /proc/meminfo | grep Cma
CmaTotal: 2097152 kB
CmaFree: 1972064 kB
Note: We print even without CONFIG_CMA, just like "nr_free_cma"; this way,
one can be sure when spotting "cma 0", that there are definetly no
CMA pages located in a zone.
[david@redhat.com: v2]
Link: https://lkml.kernel.org/r/20210128164533.18566-1-david@redhat.com
[david@redhat.com: v3]
Link: https://lkml.kernel.org/r/20210129113451.22085-1-david@redhat.com
Link: https://lkml.kernel.org/r/20210127101813.6370-3-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch adds swapcache stat for the cgroup v2. The swapcache
represents the memory that is accounted against both the memory and the
swap limit of the cgroup. The main motivation behind exposing the
swapcache stat is for enabling users to gracefully migrate from cgroup
v1's memsw counter to cgroup v2's memory and swap counters.
Cgroup v1's memsw limit allows users to limit the memory+swap usage of a
workload but without control on the exact proportion of memory and swap.
Cgroup v2 provides separate limits for memory and swap which enables more
control on the exact usage of memory and swap individually for the
workload.
With some little subtleties, the v1's memsw limit can be switched with the
sum of the v2's memory and swap limits. However the alternative for memsw
usage is not yet available in cgroup v2. Exposing per-cgroup swapcache
stat enables that alternative. Adding the memory usage and swap usage and
subtracting the swapcache will approximate the memsw usage. This will
help in the transparent migration of the workloads depending on memsw
usage and limit to v2' memory and swap counters.
The reasons these applications are still interested in this approximate
memsw usage are: (1) these applications are not really interested in two
separate memory and swap usage metrics. A single usage metric is more
simple to use and reason about for them.
(2) The memsw usage metric hides the underlying system's swap setup from
the applications. Applications with multiple instances running in a
datacenter with heterogeneous systems (some have swap and some don't) will
keep seeing a consistent view of their usage.
[akpm@linux-foundation.org: fix CONFIG_SWAP=n build]
Link: https://lkml.kernel.org/r/20210108155813.2914586-3-shakeelb@google.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Roman Gushchin <guro@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Muchun Song <songmuchun@bytedance.com>
Cc: Yang Shi <shy828301@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently we use struct per_cpu_nodestat to cache the vmstat counters,
which leads to inaccurate statistics especially THP vmstat counters. In
the systems with hundreds of processors it can be GBs of memory. For
example, for a 96 CPUs system, the threshold is the maximum number of 125.
And the per cpu counters can cache 23.4375 GB in total.
The THP page is already a form of batched addition (it will add 512 worth
of memory in one go) so skipping the batching seems like sensible.
Although every THP stats update overflows the per-cpu counter, resorting
to atomic global updates. But it can make the statistics more accuracy
for the THP vmstat counters.
So we convert the NR_ANON_THPS account to pages. This patch is consistent
with 8f182270df ("mm/swap.c: flush lru pvecs on compound page arrival").
Doing this also can make the unit of vmstat counters more unified.
Finally, the unit of the vmstat counters are pages, kB and bytes. The
B/KB suffix can tell us that the unit is bytes or kB. The rest which is
without suffix are pages.
Link: https://lkml.kernel.org/r/20201228164110.2838-3-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Rafael. J. Wysocki <rafael@kernel.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Sami Tolvanen <samitolvanen@google.com>
Cc: Feng Tang <feng.tang@intel.com>
Cc: NeilBrown <neilb@suse.de>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pankaj Gupta <pankaj.gupta@cloud.ionos.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For many workloads, pagetable consumption is significant and it makes
sense to expose it in the memory.stat for the memory cgroups. However at
the moment, the pagetables are accounted per-zone. Converting them to
per-node and using the right interface will correctly account for the
memory cgroups as well.
[akpm@linux-foundation.org: export __mod_lruvec_page_state to modules for arch/mips/kvm/]
Link: https://lkml.kernel.org/r/20201130212541.2781790-3-shakeelb@google.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Roman Gushchin <guro@fb.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>
Steps on the way to 5.10-rc1
Resolves conflicts in:
fs/userfaultfd.c
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ie3fe3c818f1f6565cfd4fa551de72d2b72ef60af
Merge emailed patches from Peter Xu:
"This is a small series that I picked up from Linus's suggestion to
simplify cow handling (and also make it more strict) by checking
against page refcounts rather than mapcounts.
This makes uffd-wp work again (verified by running upmapsort)"
Note: this is horrendously bad timing, and making this kind of
fundamental vm change after -rc3 is not at all how things should work.
The saving grace is that it really is a a nice simplification:
8 files changed, 29 insertions(+), 120 deletions(-)
The reason for the bad timing is that it turns out that commit
17839856fd ("gup: document and work around 'COW can break either way'
issue" broke not just UFFD functionality (as Peter noticed), but Mikulas
Patocka also reports that it caused issues for strace when running in a
DAX environment with ext4 on a persistent memory setup.
And we can't just revert that commit without re-introducing the original
issue that is a potential security hole, so making COW stricter (and in
the process much simpler) is a step to then undoing the forced COW that
broke other uses.
Link: https://lore.kernel.org/lkml/alpine.LRH.2.02.2009031328040.6929@file01.intranet.prod.int.rdu2.redhat.com/
* emailed patches from Peter Xu <peterx@redhat.com>:
mm: Add PGREUSE counter
mm/gup: Remove enfornced COW mechanism
mm/ksm: Remove reuse_ksm_page()
mm: do_wp_page() simplification
For some applications, we need to allocate almost all memory as hugepages.
However, on a running system, higher-order allocations can fail if the
memory is fragmented. Linux kernel currently does on-demand compaction as
we request more hugepages, but this style of compaction incurs very high
latency. Experiments with one-time full memory compaction (followed by
hugepage allocations) show that kernel is able to restore a highly
fragmented memory state to a fairly compacted memory state within <1 sec
for a 32G system. Such data suggests that a more proactive compaction can
help us allocate a large fraction of memory as hugepages keeping
allocation latencies low.
For a more proactive compaction, the approach taken here is to define a
new sysctl called 'vm.compaction_proactiveness' which dictates bounds for
external fragmentation which kcompactd tries to maintain.
The tunable takes a value in range [0, 100], with a default of 20.
Note that a previous version of this patch [1] was found to introduce too
many tunables (per-order extfrag{low, high}), but this one reduces them to
just one sysctl. Also, the new tunable is an opaque value instead of
asking for specific bounds of "external fragmentation", which would have
been difficult to estimate. The internal interpretation of this opaque
value allows for future fine-tuning.
Currently, we use a simple translation from this tunable to [low, high]
"fragmentation score" thresholds (low=100-proactiveness, high=low+10%).
The score for a node is defined as weighted mean of per-zone external
fragmentation. A zone's present_pages determines its weight.
To periodically check per-node score, we reuse per-node kcompactd threads,
which are woken up every 500 milliseconds to check the same. If a node's
score exceeds its high threshold (as derived from user-provided
proactiveness value), proactive compaction is started until its score
reaches its low threshold value. By default, proactiveness is set to 20,
which implies threshold values of low=80 and high=90.
This patch is largely based on ideas from Michal Hocko [2]. See also the
LWN article [3].
Performance data
================
System: x64_64, 1T RAM, 80 CPU threads.
Kernel: 5.6.0-rc3 + this patch
echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/enabled
echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/defrag
Before starting the driver, the system was fragmented from a userspace
program that allocates all memory and then for each 2M aligned section,
frees 3/4 of base pages using munmap. The workload is mainly anonymous
userspace pages, which are easy to move around. I intentionally avoided
unmovable pages in this test to see how much latency we incur when
hugepage allocations hit direct compaction.
1. Kernel hugepage allocation latencies
With the system in such a fragmented state, a kernel driver then allocates
as many hugepages as possible and measures allocation latency:
(all latency values are in microseconds)
- With vanilla 5.6.0-rc3
percentile latency
–––––––––– –––––––
5 7894
10 9496
25 12561
30 15295
40 18244
50 21229
60 27556
75 30147
80 31047
90 32859
95 33799
Total 2M hugepages allocated = 383859 (749G worth of hugepages out of 762G
total free => 98% of free memory could be allocated as hugepages)
- With 5.6.0-rc3 + this patch, with proactiveness=20
sysctl -w vm.compaction_proactiveness=20
percentile latency
–––––––––– –––––––
5 2
10 2
25 3
30 3
40 3
50 4
60 4
75 4
80 4
90 5
95 429
Total 2M hugepages allocated = 384105 (750G worth of hugepages out of 762G
total free => 98% of free memory could be allocated as hugepages)
2. JAVA heap allocation
In this test, we first fragment memory using the same method as for (1).
Then, we start a Java process with a heap size set to 700G and request the
heap to be allocated with THP hugepages. We also set THP to madvise to
allow hugepage backing of this heap.
/usr/bin/time
java -Xms700G -Xmx700G -XX:+UseTransparentHugePages -XX:+AlwaysPreTouch
The above command allocates 700G of Java heap using hugepages.
- With vanilla 5.6.0-rc3
17.39user 1666.48system 27:37.89elapsed
- With 5.6.0-rc3 + this patch, with proactiveness=20
8.35user 194.58system 3:19.62elapsed
Elapsed time remains around 3:15, as proactiveness is further increased.
Note that proactive compaction happens throughout the runtime of these
workloads. The situation of one-time compaction, sufficient to supply
hugepages for following allocation stream, can probably happen for more
extreme proactiveness values, like 80 or 90.
In the above Java workload, proactiveness is set to 20. The test starts
with a node's score of 80 or higher, depending on the delay between the
fragmentation step and starting the benchmark, which gives more-or-less
time for the initial round of compaction. As t he benchmark consumes
hugepages, node's score quickly rises above the high threshold (90) and
proactive compaction starts again, which brings down the score to the low
threshold level (80). Repeat.
bpftrace also confirms proactive compaction running 20+ times during the
runtime of this Java benchmark. kcompactd threads consume 100% of one of
the CPUs while it tries to bring a node's score within thresholds.
Backoff behavior
================
Above workloads produce a memory state which is easy to compact. However,
if memory is filled with unmovable pages, proactive compaction should
essentially back off. To test this aspect:
- Created a kernel driver that allocates almost all memory as hugepages
followed by freeing first 3/4 of each hugepage.
- Set proactiveness=40
- Note that proactive_compact_node() is deferred maximum number of times
with HPAGE_FRAG_CHECK_INTERVAL_MSEC of wait between each check
(=> ~30 seconds between retries).
[1] https://patchwork.kernel.org/patch/11098289/
[2] https://lore.kernel.org/linux-mm/20161230131412.GI13301@dhcp22.suse.cz/
[3] https://lwn.net/Articles/817905/
Signed-off-by: Nitin Gupta <nigupta@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Oleksandr Natalenko <oleksandr@redhat.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Khalid Aziz <khalid.aziz@oracle.com>
Reviewed-by: Oleksandr Natalenko <oleksandr@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Nitin Gupta <ngupta@nitingupta.dev>
Cc: Oleksandr Natalenko <oleksandr@redhat.com>
Link: http://lkml.kernel.org/r/20200616204527.19185-1-nigupta@nvidia.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Merges along the way to 5.9-rc1
resolves conflicts in:
Documentation/ABI/testing/sysfs-class-power
drivers/power/supply/power_supply_sysfs.c
fs/crypto/inline_crypt.c
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: Ia087834f54fb4e5269d68c3c404747ceed240701
