The page allocator groups requests by migratetype to stave off
fragmentation. However, in practice this is routinely defeated by the
fact that it gives up *before* invoking reclaim and compaction - which may
well produce suitable pages. As a result, fragmentation of physical
memory is a common ongoing process in many load scenarios.
Fragmentation deteriorates compaction's ability to produce huge pages.
Depending on the lifetime of the fragmenting allocations, those effects
can be long-lasting or even permanent, requiring drastic measures like
forcible idle states or even reboots as the only reliable ways to recover
the address space for THP production.
In a kernel build test with supplemental THP pressure, the THP allocation
rate steadily declines over 15 runs:
thp_fault_alloc
61988
56474
57258
50187
52388
55409
52925
47648
43669
40621
36077
41721
36685
34641
33215
This is a hurdle in adopting THP in any environment where hosts are shared
between multiple overlapping workloads (cloud environments), and rarely
experience true idle periods. To make THP a reliable and predictable
optimization, there needs to be a stronger guarantee to avoid such
fragmentation.
Introduce defrag_mode. When enabled, reclaim/compaction is invoked to its
full extent *before* falling back. Specifically, ALLOC_NOFRAGMENT is
enforced on the allocator fastpath and the reclaiming slowpath.
For now, fallbacks are permitted to avert OOMs. There is a plan to add
defrag_mode=2 to prefer OOMs over fragmentation, but this requires
additional prep work in compaction and the reserve management to make it
ready for all possible allocation contexts.
The following test results are from a kernel build with periodic bursts of
THP allocations, over 15 runs:
vanilla defrag_mode=1
@claimer[unmovable]: 189 103
@claimer[movable]: 92 103
@claimer[reclaimable]: 207 61
@pollute[unmovable from movable]: 25 0
@pollute[unmovable from reclaimable]: 28 0
@pollute[movable from unmovable]: 38835 0
@pollute[movable from reclaimable]: 147136 0
@pollute[reclaimable from unmovable]: 178 0
@pollute[reclaimable from movable]: 33 0
@steal[unmovable from movable]: 11 0
@steal[unmovable from reclaimable]: 5 0
@steal[reclaimable from unmovable]: 107 0
@steal[reclaimable from movable]: 90 0
@steal[movable from reclaimable]: 354 0
@steal[movable from unmovable]: 130 0
Both types of polluting fallbacks are eliminated in this workload.
Interestingly, whole block conversions are reduced as well. This is
because once a block is claimed for a type, its empty space remains
available for future allocations, instead of being padded with fallbacks;
this allows the native type to group up instead of spreading out to new
blocks. The assumption in the allocator has been that pollution from
movable allocations is less harmful than from other types, since they can
be reclaimed or migrated out should the space be needed. However, since
fallbacks occur *before* reclaim/compaction is invoked, movable pollution
will still cause non-movable allocations to spread out and claim more
blocks.
Without fragmentation, THP rates hold steady with defrag_mode=1:
thp_fault_alloc
32478
20725
45045
32130
14018
21711
40791
29134
34458
45381
28305
17265
22584
28454
30850
While the downward trend is eliminated, the keen reader will of course
notice that the baseline rate is much smaller than the vanilla kernel's to
begin with. This is due to deficiencies in how reclaim and compaction are
currently driven: ALLOC_NOFRAGMENT increases the extent to which smaller
allocations are competing with THPs for pageblocks, while making no effort
themselves to reclaim or compact beyond their own request size. This
effect already exists with the current usage of ALLOC_NOFRAGMENT, but is
amplified by defrag_mode insisting on whole block stealing much more
strongly.
Subsequent patches will address defrag_mode reclaim strategy to raise the
THP success baseline above the vanilla kernel.
Link: https://lkml.kernel.org/r/20250313210647.1314586-4-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "Buddy allocator like (or non-uniform) folio split", v10.
This patchset adds a new buddy allocator like (or non-uniform) large folio
split from a order-n folio to order-m with m < n. It reduces
1. the total number of after-split folios from 2^(n-m) to n-m+1;
2. the amount of memory needed for multi-index xarray split from 2^(n/6-m/6) to
n/6-m/6, assuming XA_CHUNK_SHIFT=6;
3. keep more large folios after a split from all order-m folios to
order-(n-1) to order-m folios.
For example, to split an order-9 to order-0, folio split generates 10 (or
11 for anonymous memory) folios instead of 512, allocates 1 xa_node
instead of 8, and leaves 1 order-8, 1 order-7, ..., 1 order-1 and 2
order-0 folios (or 4 order-0 for anonymous memory) instead of 512 order-0
folios.
Instead of duplicating existing split_huge_page*() code, __folio_split()
is introduced as the shared backend code for both
split_huge_page_to_list_to_order() and folio_split(). __folio_split() can
support both uniform split and buddy allocator like (or non-uniform)
split. All existing split_huge_page*() users can be gradually converted
to use folio_split() if possible. In this patchset, I converted
truncate_inode_partial_folio() to use folio_split().
xfstests quick group passed for both tmpfs and xfs. I also
semi-replicated Hugh's test[12] and ran it without any issue for almost 24
hours.
This patch (of 8):
A preparation patch for non-uniform folio split, which always split a
folio into half iteratively, and minimal xarray entry split.
Currently, xas_split_alloc() and xas_split() always split all slots from a
multi-index entry. They cost the same number of xa_node as the
to-be-split slots. For example, to split an order-9 entry, which takes
2^(9-6)=8 slots, assuming XA_CHUNK_SHIFT is 6 (!CONFIG_BASE_SMALL), 8
xa_node are needed. Instead xas_try_split() is intended to be used
iteratively to split the order-9 entry into 2 order-8 entries, then split
one order-8 entry, based on the given index, to 2 order-7 entries, ...,
and split one order-1 entry to 2 order-0 entries. When splitting the
order-6 entry and a new xa_node is needed, xas_try_split() will try to
allocate one if possible. As a result, xas_try_split() would only need 1
xa_node instead of 8.
When a new xa_node is needed during the split, xas_try_split() can try to
allocate one but no more. -ENOMEM will be return if a node cannot be
allocated. -EINVAL will be return if a sibling node is split or cascade
split happens, where two or more new nodes are needed, and these are not
supported by xas_try_split().
xas_split_alloc() and xas_split() split an order-9 to order-0:
---------------------------------
| | | | | | | | |
| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| | | | | | | | |
---------------------------------
| | | |
------- --- --- -------
| | ... | |
V V V V
----------- ----------- ----------- -----------
| xa_node | | xa_node | ... | xa_node | | xa_node |
----------- ----------- ----------- -----------
xas_try_split() splits an order-9 to order-0:
---------------------------------
| | | | | | | | |
| 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| | | | | | | | |
---------------------------------
|
|
V
-----------
| xa_node |
-----------
Link: https://lkml.kernel.org/r/20250307174001.242794-1-ziy@nvidia.com
Link: https://lkml.kernel.org/r/20250307174001.242794-2-ziy@nvidia.com
Signed-off-by: Zi Yan <ziy@nvidia.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Kefeng Wang <wangkefeng.wang@huawei.com>
Cc: Kirill A. Shuemov <kirill.shutemov@linux.intel.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Yang Shi <yang@os.amperecomputing.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Kairui Song <kasong@tencent.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Everything is in place to stop using the per-page mapcounts in large
folios: the mapcount of tail pages will always be logically 0 (-1 value),
just like it currently is for hugetlb folios already, and the page
mapcount of the head page is either 0 (-1 value) or contains a page type
(e.g., hugetlb).
Maintaining _nr_pages_mapped without per-page mapcounts is impossible, so
that one also has to go with CONFIG_NO_PAGE_MAPCOUNT.
There are two remaining implications:
(1) Per-node, per-cgroup and per-lruvec stats of "NR_ANON_MAPPED"
("mapped anonymous memory") and "NR_FILE_MAPPED"
("mapped file memory"):
As soon as any page of the folio is mapped -- folio_mapped() -- we
now account the complete folio as mapped. Once the last page is
unmapped -- !folio_mapped() -- we account the complete folio as
unmapped.
This implies that ...
* "AnonPages" and "Mapped" in /proc/meminfo and
/sys/devices/system/node/*/meminfo
* cgroup v2: "anon" and "file_mapped" in "memory.stat" and
"memory.numa_stat"
* cgroup v1: "rss" and "mapped_file" in "memory.stat" and
"memory.numa_stat
... can now appear higher than before. But note that these folios do
consume that memory, simply not all pages are actually currently
mapped.
It's worth nothing that other accounting in the kernel (esp. cgroup
charging on allocation) is not affected by this change.
[why oh why is "anon" called "rss" in cgroup v1]
(2) Detecting partial mappings
Detecting whether anon THPs are partially mapped gets a bit more
unreliable. As long as a single MM maps such a large folio
("exclusively mapped"), we can reliably detect it. Especially before
fork() / after a short-lived child process quit, we will detect
partial mappings reliably, which is the common case.
In essence, if the average per-page mapcount in an anon THP is < 1,
we know for sure that we have a partial mapping.
However, as soon as multiple MMs are involved, we might miss detecting
partial mappings: this might be relevant with long-lived child
processes. If we have a fully-mapped anon folio before fork(), once
our child processes and our parent all unmap (zap/COW) the same pages
(but not the complete folio), we might not detect the partial mapping.
However, once the child processes quit we would detect the partial
mapping.
How relevant this case is in practice remains to be seen.
Swapout/migration will likely mitigate this.
In the future, RMAP walkers could check for that for that case
(e.g., when collecting access bits during reclaim) and simply flag
them for deferred-splitting.
Link: https://lkml.kernel.org/r/20250303163014.1128035-21-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Andy Lutomirks^H^Hski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Lance Yang <ioworker0@gmail.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: tejun heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zefan Li <lizefan.x@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Let's implement an alternative when per-page mapcounts in large folios are
no longer maintained -- soon with CONFIG_NO_PAGE_MAPCOUNT.
When computing the output for smaps / smaps_rollups, in particular when
calculating the USS (Unique Set Size) and the PSS (Proportional Set Size),
we still rely on per-page mapcounts.
To determine private vs. shared, we'll use folio_likely_mapped_shared(),
similar to how we handle PM_MMAP_EXCLUSIVE. Similarly, we might now
under-estimate the USS and count pages towards "shared" that are actually
"private" ("exclusively mapped").
When calculating the PSS, we'll now also use the average per-page mapcount
for large folios: this can result in both, an over-estimation and an
under-estimation of the PSS. The difference is not expected to matter
much in practice, but we'll have to learn as we go.
We can now provide folio_precise_page_mapcount() only with
CONFIG_PAGE_MAPCOUNT, and remove one of the last users of per-page
mapcounts when CONFIG_NO_PAGE_MAPCOUNT is enabled.
Document the new behavior.
Link: https://lkml.kernel.org/r/20250303163014.1128035-20-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Andy Lutomirks^H^Hski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Lance Yang <ioworker0@gmail.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: tejun heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zefan Li <lizefan.x@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Let's implement an alternative when per-page mapcounts in large folios are
no longer maintained -- soon with CONFIG_NO_PAGE_MAPCOUNT.
For large folios, we'll return the per-page average mapcount within the
folio, whereby we round to the closest integer when calculating the
average: however, we'll always return at least 1 if the folio is mapped.
So assuming a folio with 512 pages, the average would be:
* 0 if not pages are mapped
* 1 if there are 1 .. 767 per-page mappings
* 2 if there are 767 .. 1279 per-page mappings
...
For hugetlb folios and for large folios that are fully mapped into all
address spaces, there is no change.
We'll make use of this helper in other context next.
As an alternative, we could simply return 0 for non-hugetlb large folios,
or disable this legacy interface with CONFIG_NO_PAGE_MAPCOUNT.
But the information exposed by this interface can still be valuable, and
frequently we deal with fully-mapped large folios where the average
corresponds to the actual page mapcount. So we'll leave it like this for
now and document the new behavior.
Note: this interface is likely not very relevant for performance. If ever
required, we could try doing a rather expensive rmap walk to collect
precisely how often this folio page is mapped.
Link: https://lkml.kernel.org/r/20250303163014.1128035-17-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Andy Lutomirks^H^Hski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Lance Yang <ioworker0@gmail.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: tejun heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zefan Li <lizefan.x@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
For small folios, we traditionally use the mapcount to decide whether it
was "certainly mapped exclusively" by a single MM (mapcount == 1) or
whether it "maybe mapped shared" by multiple MMs (mapcount > 1). For
PMD-sized folios that were PMD-mapped, we were able to use a similar
mechanism (single PMD mapping), but for PTE-mapped folios and in the
future folios that span multiple PMDs, this does not work.
So we need a different mechanism to handle large folios. Let's add a new
mechanism to detect whether a large folio is "certainly mapped
exclusively", or whether it is "maybe mapped shared".
We'll use this information next to optimize CoW reuse for PTE-mapped
anonymous THP, and to convert folio_likely_mapped_shared() to
folio_maybe_mapped_shared(), independent of per-page mapcounts.
For each large folio, we'll have two slots, whereby a slot stores:
(1) an MM id: unique id assigned to each MM
(2) a per-MM mapcount
If a slot is unoccupied, it can be taken by the next MM that maps folio
page.
In addition, we'll remember the current state -- "mapped exclusively" vs.
"maybe mapped shared" -- and use a bit spinlock to sync on updates and to
reduce the total number of atomic accesses on updates. In the future, it
might be possible to squeeze a proper spinlock into "struct folio". For
now, keep it simple, as we require the whole thing with THP only, that is
incompatible with RT.
As we have to squeeze this information into the "struct folio" of even
folios of order-1 (2 pages), and we generally want to reduce the required
metadata, we'll assign each MM a unique ID that can fit into an int. In
total, we can squeeze everything into 4x int (2x long) on 64bit.
32bit support is a bit challenging, because we only have 2x long == 2x int
in order-1 folios. But we can make it work for now, because we neither
expect many MMs nor very large folios on 32bit.
We will reliably detect folios as "mapped exclusively" vs. "mapped
shared" as long as only two MMs map pages of a folio at one point in time
-- for example with fork() and short-lived child processes, or with apps
that hand over state from one instance to another.
As soon as three MMs are involved at the same time, we might detect "maybe
mapped shared" although the folio is "mapped exclusively".
Example 1:
(1) App1 faults in a (shmem/file-backed) folio page -> Tracked as MM0
(2) App2 faults in a folio page -> Tracked as MM1
(4) App1 unmaps all folio pages
-> We will detect "mapped exclusively".
Example 2:
(1) App1 faults in a (shmem/file-backed) folio page -> Tracked as MM0
(2) App2 faults in a folio page -> Tracked as MM1
(3) App3 faults in a folio page -> No slot available, tracked as "unknown"
(4) App1 and App2 unmap all folio pages
-> We will detect "maybe mapped shared".
Make use of __always_inline to keep possible performance degradation when
(un)mapping large folios to a minimum.
Note: by squeezing the two flags into the "unsigned long" that stores the
MM ids, we can use non-atomic __bit_spin_unlock() and non-atomic
setting/clearing of the "maybe mapped shared" bit, effectively not adding
any new atomics on the hot path when updating the large mapcount + new
metadata, which further helps reduce the runtime overhead in
micro-benchmarks.
Link: https://lkml.kernel.org/r/20250303163014.1128035-13-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Andy Lutomirks^H^Hski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Lance Yang <ioworker0@gmail.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: tejun heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zefan Li <lizefan.x@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "fs/proc/task_mmu: add guard region bit to pagemap".
Currently there is no means of determining whether a given page in a
mapping range is designated a guard region (as installed via madvise()
using the MADV_GUARD_INSTALL flag).
This is generally not an issue, but in some instances users may wish to
determine whether this is the case.
This series adds this ability via /proc/$pid/pagemap, updates the
documentation and adds a self test to assert that this functions
correctly.
This patch (of 2):
Currently there is no means by which users can determine whether a given
page in memory is in fact a guard region, that is having had the
MADV_GUARD_INSTALL madvise() flag applied to it.
This is intentional, as to provide this information in VMA metadata would
contradict the intent of the feature (providing a means to change fault
behaviour at a page table level rather than a VMA level), and would
require VMA metadata operations to scan page tables, which is
unacceptable.
In many cases, users have no need to reflect and determine what regions
have been designated guard regions, as it is the user who has established
them in the first place.
But in some instances, such as monitoring software, or software that
relies upon being able to ascertain the nature of mappings within a remote
process for instance, it becomes useful to be able to determine which
pages have the guard region marker applied.
This patch makes use of an unused pagemap bit (58) to provide this
information.
This patch updates the documentation at the same time as making the change
such that the implementation of the feature and the documentation of it
are tied together.
Link: https://lkml.kernel.org/r/cover.1740139449.git.lorenzo.stoakes@oracle.com
Link: https://lkml.kernel.org/r/521d99c08b975fb06a1e7201e971cc24d68196d1.1740139449.git.lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kalesh Singh <kaleshsingh@google.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Currently, CMA manages one range of physically contiguous memory.
Creation of larger CMA areas with hugetlb_cma may run in to gaps in
physical memory, so that they are not able to allocate that contiguous
physical range from memblock when creating the CMA area.
This can happen, for example, on an AMD system with > 1TB of memory, where
there will be a gap just below the 1TB (40bit DMA) line. If you have set
aside most of memory for potential hugetlb CMA allocation,
cma_declare_contiguous_nid will fail.
hugetlb_cma doesn't need the entire area to be one physically contiguous
range. It just cares about being able to get physically contiguous chunks
of a certain size (e.g. 1G), and it is fine to have the CMA area backed
by multiple physical ranges, as long as it gets 1G contiguous allocations.
Multi-range support is implemented by introducing an array of ranges,
instead of just one big one. Each range has its own bitmap. Effectively,
the allocate and release operations work as before, just per-range. So,
instead of going through one large bitmap, they now go through a number of
smaller ones.
The maximum number of supported ranges is 8, as defined in CMA_MAX_RANGES.
Since some current users of CMA expect a CMA area to just use one
physically contiguous range, only allow for multiple ranges if a new
interface, cma_declare_contiguous_nid_multi, is used. The other
interfaces will work like before, creating only CMA areas with 1 range.
cma_declare_contiguous_nid_multi works as follows, mimicking the
default "bottom-up, above 4G" reservation approach:
0) Try cma_declare_contiguous_nid, which will use only one
region. If this succeeds, return. This makes sure that for
all the cases that currently work, the behavior remains
unchanged even if the caller switches from
cma_declare_contiguous_nid to cma_declare_contiguous_nid_multi.
1) Select the largest free memblock ranges above 4G, with
a maximum number of CMA_MAX_RANGES.
2) If we did not find at most CMA_MAX_RANGES that add
up to the total size requested, return -ENOMEM.
3) Sort the selected ranges by base address.
4) Reserve them bottom-up until we get what we wanted.
Link: https://lkml.kernel.org/r/20250228182928.2645936-3-fvdl@google.com
Signed-off-by: Frank van der Linden <fvdl@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Joao Martins <joao.m.martins@oracle.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Roman Gushchin (Cruise) <roman.gushchin@linux.dev>
Cc: Usama Arif <usamaarif642@gmail.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "hugetlb/CMA improvements for large systems", v5.
On large systems, we observed some issues with hugetlb and CMA:
1) When specifying a large number of hugetlb boot pages (hugepages= on
the commandline), the kernel may run out of memory before it even gets
to HVO. For example, if you have a 3072G system, and want to use 3024
1G hugetlb pages for VMs, that should leave you plenty of space for the
hypervisor, provided you have the hugetlb vmemmap optimization (HVO)
enabled. However, since the vmemmap pages are always allocated first,
and then later in boot freed, you will actually run yourself out of
memory before you can do HVO. This means not getting all the hugetlb
pages you want, and worse, failure to boot if there is an allocation
failure in the system from which it can't recover.
2) There is a system setup where you might want to use hugetlb_cma with
a large value (say, again, 3024 out of 3072G like above), and then
lower that if system usage allows it, to make room for non-hugetlb
processes. For this, a variation of the problem above applies: the
kernel runs out of unmovable space to allocate from before you finish
boot, since your CMA area takes up all the space.
3) CMA wants to use one big contiguous area for allocations. Which
fails if you have the aforementioned 3T system with a gap in the middle
of physical memory (like the < 40bits BIOS DMA area seen on some AMD
systems). You then won't be able to set up a CMA area for one of the
NUMA nodes, leading to loss of half of your hugetlb CMA area.
4) Under the scenario mentioned in 2), when trying to grow the number
of hugetlb pages after dropping it for a while, new CMA allocations may
fail occasionally. This is not unexpected, some transient references
on pages may prevent cma_alloc from succeeding under memory pressure.
However, the hugetlb code then falls back to a normal contiguous alloc,
which may end up succeeding. This is not always desired behavior. If
you have a large CMA area, then the kernel has a restricted amount of
memory it can do unmovable allocations from (a well known issue). A
normal contiguous alloc may eat further in to this space.
To resolve these issues, do the following:
* Add hooks to the section init code to do custom initialization of
memmap pages. Hugetlb bootmem (memblock) allocated pages can then be
pre-HVOed. This avoids allocating a large number of vmemmap pages early
in boot, only to have them be freed again later, and also avoids running
out of memory as described under 1). Using these hooks for hugetlb is
optional. It requires moving hugetlb bootmem allocation to an earlier
spot by the architecture. This has been enabled on x86.
* hugetlb_cma doesn't care about the CMA area it uses being one large
contiguous range. Multiple smaller ranges are fine. The only
requirements are that the areas should be on one NUMA node, and
individual gigantic pages should be allocatable from them. So,
implement multi-range support for CMA, avoiding issue 3).
* Introduce a hugetlb_cma_only option on the commandline. This only
allows allocations from CMA for gigantic pages, if hugetlb_cma= is also
specified.
* With hugetlb_cma_only active, it also makes sense to be able to
pre-allocate gigantic hugetlb pages at boot time from the CMA area(s).
Add a rudimentary early CMA allocation interface, that just grabs a
piece of memblock-allocated space from the CMA area, which gets marked
as allocated in the CMA bitmap when the CMA area is initialized. With
this, hugepages= can be supported with hugetlb_cma=, making scenario 2)
work.
Additionally, fix some minor bugs, with one worth mentioning: since
hugetlb gigantic bootmem pages are allocated by memblock, they may span
multiple zones, as memblock doesn't (and mostly can't) know about zones.
This can cause problems. A hugetlb page spanning multiple zones is bad,
and it's worse with HVO, when the de-HVO step effectively sneakily
re-assigns pages to a different zone than originally configured, since the
tail pages all inherit the zone from the first 60 tail pages. This
condition is not common, but can be easily reproduced using ZONE_MOVABLE.
To fix this, add checks to see if gigantic bootmem pages intersect with
multiple zones, and do not use them if they do, giving them back to the
page allocator instead.
The first patch is kind of along for the ride, except that maintaining an
available_count for a CMA area is convenient for the multiple range
support.
This patch (of 27):
In addition to the number of allocations and releases, system management
software may like to be aware of the size of CMA areas, and how many pages
are available in it. This information is currently not available, so
export it in total_page and available_pages, respectively.
The name 'available_pages' was picked over 'free_pages' because 'free'
implies that the pages are unused. But they might not be, they just
haven't been used by cma_alloc
The number of available pages is tracked regardless of CONFIG_CMA_SYSFS,
allowing for a few minor shortcuts in the code, avoiding bitmap
operations.
Link: https://lkml.kernel.org/r/20250228182928.2645936-2-fvdl@google.com
Signed-off-by: Frank van der Linden <fvdl@google.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Joao Martins <joao.m.martins@oracle.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Roman Gushchin (Cruise) <roman.gushchin@linux.dev>
Cc: Usama Arif <usamaarif642@gmail.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Alexander Gordeev <agordeev@linux.ibm.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Madhavan Srinivasan <maddy@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
On what DAMON layer a DAMOS filter is handled is important to expect in
what order filters will be evaluated. Re-organize the DAMOS filter types
list on the design doc to categorize types based on the handling layer, to
let users more easily understand the handling order.
Link: https://lkml.kernel.org/r/20250218223708.53437-6-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
If an element of memory matches a DAMOS filter, filters that installed
after that get no chance to make any effect to the element. Hence in what
order DAMOS filters are handled is important, if both allow filters and
reject filters are used together.
The ordering is affected by both the installation order and which layter
the filters are handled. The design document is not clearly documenting
the latter part. Clarify it on the design doc.
Link: https://lkml.kernel.org/r/20250218223708.53437-5-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
DAMON sysfs usage doc is describing DAMOS filter type names and their
meanings in short. The design doc is providing the short meaning and
detailed descriptions, too. This is unnecessary duplicates and confuses
where to document new DAMOS filter types and features. Move the details
from usage to design doc.
Link: https://lkml.kernel.org/r/20250218223708.53437-4-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "Docs/mm/damon: misc DAMOS filters documentation fixes and
improves".
Fix and improve DAMOS filters documentation by fixing a copy-paste typo,
adding hugepage_size filter documentation on design doc, moving logic
details from usage to design, clarify DAMOS filters handling sequence
based on handling layer, and re-organizing the filters type list for
easier understanding of the handling sequence.
This patch (of 5):
The link from DAMOS filters design doc to usage doc has a typo calling
filters as watermarks. Fix it.
Link: https://lkml.kernel.org/r/20250218223708.53437-1-sj@kernel.org
Link: https://lkml.kernel.org/r/20250218223708.53437-2-sj@kernel.org
Fixes: d31f5626a0 ("Docs/mm/damon/design: add links to sections of DAMON sysfs interface usage doc")
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
For speculative lookups where a successful inc_not_zero() pins the object,
but where we still need to double check if the object acquired is indeed
the one we set out to acquire (identity check), needs this validation to
happen *after* the increment. Similarly, when a new object is initialized
and its memory might have been previously occupied by another object, all
stores to initialize the object should happen *before* refcount
initialization.
Notably SLAB_TYPESAFE_BY_RCU is one such an example when this ordering is
required for reference counting.
Add refcount_{add|inc}_not_zero_acquire() to guarantee the proper ordering
between acquiring a reference count on an object and performing the
identity check for that object.
Add refcount_set_release() to guarantee proper ordering between stores
initializing object attributes and the store initializing the refcount.
refcount_set_release() should be done after all other object attributes
are initialized. Once refcount_set_release() is called, the object should
be considered visible to other tasks even if it was not yet added into an
object collection normally used to discover it. This is because other
tasks might have discovered the object previously occupying the same
memory and after memory reuse they can succeed in taking refcount for the
new object and start using it.
Object reuse example to consider:
consumer:
obj = lookup(collection, key);
if (!refcount_inc_not_zero_acquire(&obj->ref))
return;
if (READ_ONCE(obj->key) != key) { /* identity check */
put_ref(obj);
return;
}
use(obj->value);
producer:
remove(collection, obj->key);
if (!refcount_dec_and_test(&obj->ref))
return;
obj->key = KEY_INVALID;
free(obj);
obj = malloc(); /* obj is reused */
obj->key = new_key;
obj->value = new_value;
refcount_set_release(obj->ref, 1);
add(collection, new_key, obj);
refcount_{add|inc}_not_zero_acquire() is required to prevent the following
reordering when refcount_inc_not_zero() is used instead:
consumer:
obj = lookup(collection, key);
if (READ_ONCE(obj->key) != key) { /* reordered identity check */
put_ref(obj);
return;
}
producer:
remove(collection, obj->key);
if (!refcount_dec_and_test(&obj->ref))
return;
obj->key = KEY_INVALID;
free(obj);
obj = malloc(); /* obj is reused */
obj->key = new_key;
obj->value = new_value;
refcount_set_release(obj->ref, 1);
add(collection, new_key, obj);
if (!refcount_inc_not_zero(&obj->ref))
return;
use(obj->value); /* USING WRONG OBJECT */
refcount_set_release() is required to prevent the following reordering
when refcount_set() is used instead:
consumer:
obj = lookup(collection, key);
producer:
remove(collection, obj->key);
if (!refcount_dec_and_test(&obj->ref))
return;
obj->key = KEY_INVALID;
free(obj);
obj = malloc(); /* obj is reused */
obj->key = new_key; /* new_key == old_key */
refcount_set(obj->ref, 1);
if (!refcount_inc_not_zero_acquire(&obj->ref))
return;
if (READ_ONCE(obj->key) != key) { /* pass since new_key == old_key */
put_ref(obj);
return;
}
use(obj->value); /* USING STALE obj->value */
obj->value = new_value; /* reordered store */
add(collection, key, obj);
[surenb@google.com: fix title underlines in refcount-vs-atomic.rst]
Link: https://lkml.kernel.org/r/20250217161645.3137927-1-surenb@google.com
Link: https://lkml.kernel.org/r/20250213224655.1680278-11-surenb@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz> [slab]
Tested-by: Shivank Garg <shivankg@amd.com>
Link: https://lkml.kernel.org/r/5e19ec93-8307-47c2-bb13-3ddf7150624e@amd.com
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <will@kernel.org>
Cc: Paul E. McKenney <paulmck@kernel.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: David Hildenbrand <david@redhat.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Hugh Dickins <hughd@google.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Klara Modin <klarasmodin@gmail.com>
Cc: Liam R. Howlett <Liam.Howlett@Oracle.com>
Cc: Lokesh Gidra <lokeshgidra@google.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Mateusz Guzik <mjguzik@gmail.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Minchan Kim <minchan@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Pasha Tatashin <pasha.tatashin@soleen.com>
Cc: Peter Xu <peterx@redhat.com>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Sourav Panda <souravpanda@google.com>
Cc: Wei Yang <richard.weiyang@gmail.com>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Pull Kbuild fixes from Masahiro Yamada:
- Use the specified $(LD) when building userprogs with Clang
- Pass the correct target triple when compile-testing UAPI headers
with Clang
- Fix pacman-pkg build error with KBUILD_OUTPUT
* tag 'kbuild-fixes-v6.14-3' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
kbuild: install-extmod-build: Fix build when specifying KBUILD_OUTPUT
docs: Kconfig: fix defconfig description
kbuild: hdrcheck: fix cross build with clang
kbuild: userprogs: use correct lld when linking through clang
Pull char/misc/IIO driver fixes from Greg KH:
"Here are a number of misc and char and iio driver fixes that have been
sitting in my tree for way too long. They contain:
- iio driver fixes for reported issues
- regression fix for rtsx_usb card reader
- mei and mhi driver fixes
- small virt driver fixes
- ntsync permissions fix
- other tiny driver fixes for reported problems.
All of these have been in linux-next for quite a while with no
reported issues"
* tag 'char-misc-6.14-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (30 commits)
Revert "drivers/card_reader/rtsx_usb: Restore interrupt based detection"
ntsync: Check wait count based on byte size.
bus: simple-pm-bus: fix forced runtime PM use
char: misc: deallocate static minor in error path
eeprom: digsy_mtc: Make GPIO lookup table match the device
drivers: virt: acrn: hsm: Use kzalloc to avoid info leak in pmcmd_ioctl
binderfs: fix use-after-free in binder_devices
slimbus: messaging: Free transaction ID in delayed interrupt scenario
vbox: add HAS_IOPORT dependency
cdx: Fix possible UAF error in driver_override_show()
intel_th: pci: Add Panther Lake-P/U support
intel_th: pci: Add Panther Lake-H support
intel_th: pci: Add Arrow Lake support
intel_th: msu: Fix less trivial kernel-doc warnings
intel_th: msu: Fix kernel-doc warnings
MAINTAINERS: change maintainer for FSI
ntsync: Set the permissions to be 0666
bus: mhi: host: pci_generic: Use pci_try_reset_function() to avoid deadlock
mei: vsc: Use "wakeuphostint" when getting the host wakeup GPIO
mei: me: add panther lake P DID
...
Pull core dumping fix from Kees Cook:
- Only sort VMAs when core_sort_vma sysctl is set
* tag 'execve-v6.14-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
coredump: Only sort VMAs when core_sort_vma sysctl is set
Pull vfs fixes from Christian Brauner:
- Fix spelling mistakes in idmappings.rst
- Fix RCU warnings in override_creds()/revert_creds()
- Create new pid namespaces with default limit now that pid_max is
namespaced
* tag 'vfs-6.14-rc6.fixes' of gitolite.kernel.org:pub/scm/linux/kernel/git/vfs/vfs:
pid: Do not set pid_max in new pid namespaces
doc: correcting two prefix errors in idmappings.rst
cred: Fix RCU warnings in override/revert_creds
By default fair_server dl_server allocates 5% of the bandwidth to the root
domain. Due to this writing any value less than 5% fails due to -EBUSY:
$ cat /proc/sys/kernel/sched_rt_period_us
1000000
$ echo 49999 > /proc/sys/kernel/sched_rt_runtime_us
-bash: echo: write error: Device or resource busy
$ echo 50000 > /proc/sys/kernel/sched_rt_runtime_us
$
Since the sched_rt_runtime_us allows -1 as the minimum, put this
restriction in the documentation.
One should check average of runtime/period in
/sys/kernel/debug/sched/fair_server/cpuX/* for exact value.
Signed-off-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Juri Lelli <juri.lelli@redhat.com>
Link: https://lore.kernel.org/r/20250306052954.452005-3-sshegde@linux.ibm.com
Commit 2a86f66121 ("kbuild: use KBUILD_DEFCONFIG as the fallback
for DEFCONFIG_LIST") removed arch/$ARCH/defconfig; however,
the document has not been updated to reflect this change yet.
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Pull landlock fixes from Mickaël Salaün:
"Fixes to TCP socket identification, documentation, and tests"
* tag 'landlock-6.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/mic/linux:
selftests/landlock: Add binaries to .gitignore
selftests/landlock: Test that MPTCP actions are not restricted
selftests/landlock: Test TCP accesses with protocol=IPPROTO_TCP
landlock: Fix non-TCP sockets restriction
landlock: Minor typo and grammar fixes in IPC scoping documentation
landlock: Fix grammar error
selftests/landlock: Enable the new CONFIG_AF_UNIX_OOB
Pull powerpc fix from Madhavan Srinivasan:
- Fix for cross-reference in documentation and deprecation warning
Thanks to Andrew Donnellan and Bagas Sanjaya.
* tag 'powerpc-6.14-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
cxl: Fix cross-reference in documentation and add deprecation warning
