| Age | Commit message (Collapse) | Author |
|---|
|
Using SWAPPINESS_ANON_ONLY instead of MAX_SWAPPINESS + 1 to indicate
reclaiming only from anonymous pages makes the code more readable and
explicit.
Add some comment in the SWAPPINESS_ANON_ONLY context.
Link: https://lkml.kernel.org/r/529db7ae6098ee712b81e4df290622e4e64ac50c.1745225696.git.hezhongkun.hzk@bytedance.com
Signed-off-by: Zhongkun He <hezhongkun.hzk@bytedance.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Yosry Ahmed <yosry.ahmed@linux.dev>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The MGLRU already supports reclaiming only from anonymous memory via the
/sys/kernel/debug/lru_gen interface. Now, memory.reclaim also supports
the swappiness=max parameter to enable reclaiming solely from anonymous
memory. To unify the semantics of proactive reclaiming from anonymous
folios, the max parameter is introduced.
[hezhongkun.hzk@bytedance.com: use strcmp instead of strncmp, if swappiness is not set, use the default value]
Link: https://lkml.kernel.org/r/20250507071057.3184240-1-hezhongkun.hzk@bytedance.com
[akpm@linux-foundation.org: tweak coding style]
Link: https://lkml.kernel.org/r/65181f7745d657d664d833c26d8a94cae40538b9.1745225696.git.hezhongkun.hzk@bytedance.com
Signed-off-by: Zhongkun He <hezhongkun.hzk@bytedance.com>
Acked-by: Muchun Song <muchun.song@linux.dev>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Yosry Ahmed <yosry.ahmed@linux.dev>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Add more comments for cache_trim_mode, and the annotations provided by
Johannes Weiner in [1].
Link: https://lore.kernel.org/all/20250314141833.GA1316033@cmpxchg.org/
Link: https://lkml.kernel.org/r/4baad87ba637f1e6f666e9b99b3fdcb7ab39171b.1745225696.git.hezhongkun.hzk@bytedance.com
Signed-off-by: Zhongkun He <hezhongkun.hzk@bytedance.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Yosry Ahmed <yosry.ahmed@linux.dev>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "add max arg to swappiness in memory.reclaim and lru_gen", v4.
This patchset adds max arg to swappiness in memory.reclaim and lru_gen for
anon only proactive memory reclaim.
With commit <68cd9050d871> ("mm: add swappiness= arg to memory.reclaim")
we can submit an additional swappiness=<val> argument to memory.reclaim.
It is very useful because we can dynamically adjust the reclamation ratio
based on the anonymous folios and file folios of each cgroup. For
example,when swappiness is set to 0, we only reclaim from file folios.
But we can not relciam memory just from anon folios.
This patchset introduces a new macro, SWAPPINESS_ANON_ONLY, defined as
MAX_SWAPPINESS + 1, represent the max arg semantics. It specifically
indicates that reclamation should occur only from anonymous pages.
Patch 1 adds swappiness=max arg to memory.reclaim suggested-by: Yosry
Ahmed
Patch 2 add more comments for cache_trim_mode from Johannes Weiner in [1].
Patch 3 add max arg to lru_gen for proactive memory reclaim in MGLRU. The
MGLRU already supports reclaiming exclusively from anonymous pages. This
patch formalizes that behavior by introducing a max parameter to represent
the corresponding semantics.
Patch 4 using SWAPPINESS_ANON_ONLY in MGLRU Using SWAPPINESS_ANON_ONLY
instead of MAX_SWAPPINESS + 1 to indicate reclaiming only from anonymous
pages makes the code more readable and explicit
Here is the previous discussion:
https://lore.kernel.org/all/20250314033350.1156370-1-hezhongkun.hzk@bytedance.com/
https://lore.kernel.org/all/20250312094337.2296278-1-hezhongkun.hzk@bytedance.com/
https://lore.kernel.org/all/20250318135330.3358345-1-hezhongkun.hzk@bytedance.com/
This patch (of 4):
With commit <68cd9050d871> ("mm: add swappiness= arg to memory.reclaim")
we can submit an additional swappiness=<val> argument to memory.reclaim.
It is very useful because we can dynamically adjust the reclamation ratio
based on the anonymous folios and file folios of each cgroup. For
example,when swappiness is set to 0, we only reclaim from file folios.
However,we have also encountered a new issue: when swappiness is set to
the MAX_SWAPPINESS, it may still only reclaim file folios.
So, we hope to add a new arg 'swappiness=max' in memory.reclaim where
proactive memory reclaim only reclaims from anonymous folios when
swappiness is set to max. The swappiness semantics from a user
perspective remain unchanged.
For example, something like this:
echo "2M swappiness=max" > /sys/fs/cgroup/memory.reclaim
will perform reclaim on the rootcg with a swappiness setting of 'max' (a
new mode) regardless of the file folios. Users have a more comprehensive
view of the application's memory distribution because there are many
metrics available. For example, if we find that a certain cgroup has a
large number of inactive anon folios, we can reclaim only those and skip
file folios, because with the zram/zswap, the IO tradeoff that
cache_trim_mode or other file first logic is making doesn't hold - file
refaults will cause IO, whereas anon decompression will not.
With this patch, the swappiness argument of memory.reclaim has a new
mode 'max', means reclaiming just from anonymous folios both in traditional
LRU and MGLRU.
Link: https://lkml.kernel.org/r/cover.1745225696.git.hezhongkun.hzk@bytedance.com
Link: https://lore.kernel.org/all/20250314141833.GA1316033@cmpxchg.org/ [1]
Link: https://lkml.kernel.org/r/519e12b9b1f8c31a01e228c8b4b91a2419684f77.1745225696.git.hezhongkun.hzk@bytedance.com
Signed-off-by: Zhongkun He <hezhongkun.hzk@bytedance.com>
Suggested-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Acked-by: Muchun Song <muchun.song@linux.dev>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Setting the max and high limits can trigger synchronous reclaim and/or
oom-kill if the usage is higher than the given limit. This behavior is
fine for newly created cgroups but it can cause issues for the node
controller while setting limits for existing cgroups.
In our production multi-tenant and overcommitted environment, we are
seeing priority inversion when the node controller dynamically adjusts the
limits of running jobs of different priorities. Based on the system
situation, the node controller may reduce the limits of lower priority
jobs and increase the limits of higher priority jobs. However we are
seeing node controller getting stuck for long period of time while
reclaiming from lower priority jobs while setting their limits and also
spends a lot of its own CPU.
One of the workaround we are trying is to fork a new process which sets
the limit of the lower priority job along with setting an alarm to get
itself killed if it get stuck in the reclaim for lower priority job.
However we are finding it very unreliable and costly. Either we need a
good enough time buffer for the alarm to be delivered after setting limit
and potentialy spend a lot of CPU in the reclaim or be unreliable in
setting the limit for much shorter but cheaper (less reclaim) alarms.
Let's introduce new limit setting option which does not trigger reclaim
and/or oom-kill and let the processes in the target cgroup to trigger
reclaim and/or throttling and/or oom-kill in their next charge request.
This will make the node controller on multi-tenant overcommitted
environment much more reliable.
Explanation from Johannes on side-effects of O_NONBLOCK limit change:
It's usually the allocating tasks inside the group bearing the cost of
limit enforcement and reclaim. This allows a (privileged) updater from
outside the group to keep that cost in there - instead of having to
help, from a context that doesn't necessarily make sense.
I suppose the tradeoff with that - and the reason why this was doing
sync reclaim in the first place - is that, if the group is idle and
not trying to allocate more, it can take indefinitely for the new
limit to actually be met.
It should be okay in most scenarios in practice. As the capacity is
reallocated from group A to B, B will exert pressure on A once it
tries to claim it and thereby shrink it down. If A is idle, that
shouldn't be hard. If A is running, it's likely to fault/allocate
soon-ish and then join the effort.
It does leave a (malicious) corner case where A is just busy-hitting
its memory to interfere with the clawback. This is comparable to
reclaiming memory.low overage from the outside, though, which is an
acceptable risk. Users of O_NONBLOCK just need to be aware.
Link: https://lkml.kernel.org/r/20250419183545.1982187-1-shakeel.butt@linux.dev
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Michal Koutný <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Tejun Heo <tj@kernel.org>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Yosry Ahmed <yosry.ahmed@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Hugetlb boot allocation has used online nodes for allocation since commit
de55996d7188 ("mm/hugetlb: use online nodes for bootmem allocation").
This was needed to be able to do the allocations earlier in boot, before
N_MEMORY was set.
This might lead to a different distribution of gigantic hugepages across
NUMA nodes if there are memoryless nodes in the system.
What happens is that the memoryless nodes are tried, but then the memblock
allocation fails and falls back, which usually means that the node that
has the highest physical address available will be used (top-down
allocation). While this will end up getting the same number of hugetlb
pages, they might not be be distributed the same way. The fallback for
each memoryless node might not end up coming from the same node as the
successful round-robin allocation from N_MEMORY nodes.
While administrators that rely on having a specific number of hugepages
per node should use the hugepages=N:X syntax, it's better not to change
the old behavior for the plain hugepages=N case.
To do this, construct a nodemask for hugetlb bootmem purposes only,
containing nodes that have memory. Then use that for round-robin bootmem
allocations.
This saves some cycles, and the added advantage here is that hugetlb_cma
can use it too, avoiding the older issue of pointless attempts to create a
CMA area for memoryless nodes (which will also cause the per-node CMA area
size to be too small).
Link: https://lkml.kernel.org/r/20250402205613.3086864-1-fvdl@google.com
Fixes: de55996d7188 ("mm/hugetlb: use online nodes for bootmem allocation")
Signed-off-by: Frank van der Linden <fvdl@google.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Luiz Capitulino <luizcap@redhat.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Muchun Song <muchun.song@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When tracing mem_cgroup_per_node allocations with kmalloc ftrace:
kmalloc: call_site=mem_cgroup_css_alloc+0x1d8/0x5b4 ptr=00000000d798700c
bytes_req=2896 bytes_alloc=4096 gfp_flags=GFP_KERNEL|__GFP_ZERO node=0
accounted=false
This reveals the slab allocator provides 4096B chunks for 2896B
mem_cgroup_per_node due to:
1. The slab allocator predefines bucket sizes from 64B to 8096B
2. The mem_cgroup allocation size (2312B) falls between the 2KB and 4KB
slabs
3. The allocator rounds up to the nearest larger slab (4KB), resulting in
~1KB wasted memory per memcg alloc - per node.
This patch introduces a dedicated kmem_cache for mem_cgroup structs,
achieving precise memory allocation. Post-patch ftrace verification shows:
kmem_cache_alloc: call_site=mem_cgroup_css_alloc+0x1b8/0x5d4
ptr=000000002989e63a bytes_req=2896 bytes_alloc=2944
gfp_flags=GFP_KERNEL|__GFP_ZERO node=0 accounted=false
Each mem_cgroup_per_node alloc 2944bytes(include hw cacheline align),
compare to 4096, it avoid waste.
Link: https://lkml.kernel.org/r/20250425031935.76411-4-link@vivo.com
Signed-off-by: Huan Yang <link@vivo.com>
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Francesco Valla <francesco@valla.it>
Cc: guoweikang <guoweikang.kernel@gmail.com>
Cc: Huang Shijie <shijie@os.amperecomputing.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Raul E Rangel <rrangel@chromium.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When tracing mem_cgroup_alloc() with kmalloc ftrace, we observe:
kmalloc: call_site=mem_cgroup_css_alloc+0xd8/0x5b4 ptr=000000003e4c3799
bytes_req=2312 bytes_alloc=4096 gfp_flags=GFP_KERNEL|__GFP_ZERO node=-1
accounted=false
The output indicates that while allocating mem_cgroup struct (2312 bytes),
the slab allocator actually provides 4096-byte chunks. This occurs because:
1. The slab allocator predefines bucket sizes from 64B to 8096B
2. The mem_cgroup allocation size (2312B) falls between the 2KB and 4KB
slabs
3. The allocator rounds up to the nearest larger slab (4KB), resulting in
~1KB wasted memory per allocation
This patch introduces a dedicated kmem_cache for mem_cgroup structs,
achieving precise memory allocation. Post-patch ftrace verification shows:
kmem_cache_alloc: call_site=mem_cgroup_css_alloc+0xbc/0x5d4
ptr=00000000695c1806 bytes_req=2312 bytes_alloc=2368
gfp_flags=GFP_KERNEL|__GFP_ZERO node=-1 accounted=false
Each memcg alloc offer 2368bytes(include hw cacheline align), compare to
4096, avoid waste.
Link: https://lkml.kernel.org/r/20250425031935.76411-3-link@vivo.com
Signed-off-by: Huan Yang <link@vivo.com>
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Francesco Valla <francesco@valla.it>
Cc: guoweikang <guoweikang.kernel@gmail.com>
Cc: Huang Shijie <shijie@os.amperecomputing.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Raul E Rangel <rrangel@chromium.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "Use kmem_cache for memcg alloc", v3.
(willy tldr: "you've gone from allocating 8 objects per 32KiB to
allocating 13 objects per 32KiB, a 62% improvement in memory consumption"
[1])
The mem_cgroup_alloc function creates mem_cgroup struct and it's
associated structures including mem_cgroup_per_node. Through detailed
analysis on our test machine (Arm64, 16GB RAM, 6.6 kernel, 1 NUMA node,
memcgv2 with nokmem,nosocket,cgroup_disable=pressure), we can observe the
memory allocation for these structures using the following shell commands:
# Enable tracing
echo 1 > /sys/kernel/tracing/events/kmem/kmalloc/enable
echo 1 > /sys/kernel/tracing/tracing_on
cat /sys/kernel/tracing/trace_pipe | grep kmalloc | grep mem_cgroup
# Trigger allocation if cgroup subtree do not enable memcg
echo +memory > /sys/fs/cgroup/cgroup.subtree_control
Ftrace Output:
# mem_cgroup struct allocation
sh-6312 [000] ..... 58015.698365: kmalloc:
call_site=mem_cgroup_css_alloc+0xd8/0x5b4
ptr=000000003e4c3799 bytes_req=2312 bytes_alloc=4096
gfp_flags=GFP_KERNEL|__GFP_ZERO node=-1 accounted=false
# mem_cgroup_per_node allocation
sh-6312 [000] ..... 58015.698389: kmalloc:
call_site=mem_cgroup_css_alloc+0x1d8/0x5b4
ptr=00000000d798700c bytes_req=2896 bytes_alloc=4096
gfp_flags=GFP_KERNEL|__GFP_ZERO node=0 accounted=false
Key Observations:
1. Both structures use kmalloc with requested sizes between 2KB-4KB
2. Allocation alignment forces 4KB slab usage due to pre-defined sizes
(64B, 128B,..., 2KB, 4KB, 8KB)
3. Memory waste per memcg instance:
Base struct: 4096 - 2312 = 1784 bytes
Per-node struct: 4096 - 2896 = 1200 bytes
Total waste: 2984 bytes (1-node system)
NUMA scaling: (1200 + 8) * nr_node_ids bytes
So, it's a little waste.
This patchset introduces dedicated kmem_cache:
Patch2 - mem_cgroup kmem_cache - memcg_cachep
Patch3 - mem_cgroup_per_node kmem_cache - memcg_pn_cachep
The benefits of this change can be observed with the following tracing
commands:
# Enable tracing
echo 1 > /sys/kernel/tracing/events/kmem/kmem_cache_alloc/enable
echo 1 > /sys/kernel/tracing/tracing_on
cat /sys/kernel/tracing/trace_pipe | grep kmem_cache_alloc | grep mem_cgroup
# In another terminal:
echo +memory > /sys/fs/cgroup/cgroup.subtree_control
The output might now look like this:
# mem_cgroup struct allocation
sh-9827 [000] ..... 289.513598: kmem_cache_alloc:
call_site=mem_cgroup_css_alloc+0xbc/0x5d4 ptr=00000000695c1806
bytes_req=2312 bytes_alloc=2368 gfp_flags=GFP_KERNEL|__GFP_ZERO node=-1
accounted=false
# mem_cgroup_per_node allocation
sh-9827 [000] ..... 289.513602: kmem_cache_alloc:
call_site=mem_cgroup_css_alloc+0x1b8/0x5d4 ptr=000000002989e63a
bytes_req=2896 bytes_alloc=2944 gfp_flags=GFP_KERNEL|__GFP_ZERO node=0
accounted=false
This indicates that the `mem_cgroup` struct now requests 2312 bytes and is
allocated 2368 bytes, while `mem_cgroup_per_node` requests 2896 bytes and
is allocated 2944 bytes. The slight increase in allocated size is due to
`SLAB_HWCACHE_ALIGN` in the `kmem_cache`.
Without `SLAB_HWCACHE_ALIGN`, the allocation might appear as:
# mem_cgroup struct allocation
sh-9269 [003] ..... 80.396366: kmem_cache_alloc:
call_site=mem_cgroup_css_alloc+0xbc/0x5d4 ptr=000000005b12b475
bytes_req=2312 bytes_alloc=2312 gfp_flags=GFP_KERNEL|__GFP_ZERO node=-1
accounted=false
# mem_cgroup_per_node allocation
sh-9269 [003] ..... 80.396411: kmem_cache_alloc:
call_site=mem_cgroup_css_alloc+0x1b8/0x5d4 ptr=00000000f347adc6
bytes_req=2896 bytes_alloc=2896 gfp_flags=GFP_KERNEL|__GFP_ZERO node=0
accounted=false
While the `bytes_alloc` now matches the `bytes_req`, this patchset
defaults to using `SLAB_HWCACHE_ALIGN` as it is generally considered more
beneficial for performance. Please let me know if there are any issues or
if I've misunderstood anything.
This patchset also move mem_cgroup_init ahead of cgroup_init() due to
cgroup_init() will allocate root_mem_cgroup, but each initcall invoke
after cgroup_init, so if each kmem_cache do not prepare, we need testing
NULL before use it.
This patch (of 3):
When cgroup_init() creates root_mem_cgroup through css_alloc callback,
some critical resources might not be fully initialized, forcing later
operations to perform conditional checks for resource availability.
This patch move mem_cgroup_init() to address the init order, it invoke
before cgroup_init, so, compare to subsys_initcall, it can use to prepare
some key resources before root_mem_cgroup alloc.
Link: https://lkml.kernel.org/r/aAsRCj-niMMTtmK8@casper.infradead.org [1]
Link: https://lkml.kernel.org/r/20250425031935.76411-1-link@vivo.com
Link: https://lkml.kernel.org/r/20250425031935.76411-2-link@vivo.com
Signed-off-by: Huan Yang <link@vivo.com>
Suggested-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Francesco Valla <francesco@valla.it>
Cc: guoweikang <guoweikang.kernel@gmail.com>
Cc: Huang Shijie <shijie@os.amperecomputing.com>
Cc: KP Singh <kpsingh@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: Petr Mladek <pmladek@suse.com>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Raul E Rangel <rrangel@chromium.org>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Since the previous commit "mm/huge_memory: Adjust try_to_migrate_one() and
split_huge_pmd_locked()" has simplified the logic by leveraging the folio
verification in page_vma_mapped_walk(), this patch removes the unnecessary
folio pointers passing.
Link: https://lkml.kernel.org/r/20250425103859.825879-3-gavinguo@igalia.com
Link: https://lore.kernel.org/all/98d1d195-7821-4627-b518-83103ade56c0@redhat.com/
Link: https://lore.kernel.org/all/91599a3c-e69e-4d79-bac5-5013c96203d7@redhat.com/
Signed-off-by: Gavin Guo <gavinguo@igalia.com>
Suggested-by: David Hildenbrand <david@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Florent Revest <revest@google.com>
Cc: Gavin Shan <gshan@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "Clean up split_huge_pmd_locked() and remove unnecessary
folio pointers", v2.
The patch series enhances the folio verification by leveraging the
existing page_vma_mapped_walk() mechanism and removing redundant folio
pointer passing.
This patch (of 2):
split_huge_pmd_locked() currently performs redundant checks for migration
entries and folio validation that are already handled by the
page_vma_mapped_walk mechanism in try_to_migrate_one.
Specifically, page_vma_mapped_walk already ensures that:
- The folio is properly mapped in the given VMA area
- pmd_trans_huge, pmd_devmap, and migration entry validation are
performed
To leverage page_vma_mapped_walk's work, moving TTU_SPLIT_HUGE_PMD
handling to the while loop checking and removing these duplicate checks
from split_huge_pmd_locked.
Link: https://lkml.kernel.org/r/20250425103859.825879-1-gavinguo@igalia.com
Link: https://lkml.kernel.org/r/20250425103859.825879-2-gavinguo@igalia.com
Link: https://lore.kernel.org/all/98d1d195-7821-4627-b518-83103ade56c0@redhat.com/
Link: https://lore.kernel.org/all/91599a3c-e69e-4d79-bac5-5013c96203d7@redhat.com/
Signed-off-by: Gavin Guo <gavinguo@igalia.com>
Suggested-by: David Hildenbrand <david@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Florent Revest <revest@google.com>
Cc: Gavin Shan <gshan@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
It is possible for a reclaimer to cause demotions of an lruvec belonging
to a cgroup with cpuset.mems set to exclude some nodes. Attempt to apply
this limitation based on the lruvec's memcg and prevent demotion.
Notably, this may still allow demotion of shared libraries or any memory
first instantiated in another cgroup. This means cpusets still cannot
cannot guarantee complete isolation when demotion is enabled, and the docs
have been updated to reflect this.
This is useful for isolating workloads on a multi-tenant system from
certain classes of memory more consistently - with the noted exceptions.
Note on locking:
The cgroup_get_e_css reference protects the css->effective_mems, and calls
of this interface would be subject to the same race conditions associated
with a non-atomic access to cs->effective_mems.
So while this interface cannot make strong guarantees of correctness, it
can therefore avoid taking a global or rcu_read_lock for performance.
Link: https://lkml.kernel.org/r/20250424202806.52632-3-gourry@gourry.net
Signed-off-by: Gregory Price <gourry@gourry.net>
Suggested-by: Shakeel Butt <shakeel.butt@linux.dev>
Suggested-by: Waiman Long <longman@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>
Reviewed-by: Waiman Long <longman@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michal Koutný <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "vmscan: enforce mems_effective during demotion", v5.
Change reclaim to respect cpuset.mems_effective during demotion when
possible. Presently, reclaim explicitly ignores cpuset.mems_effective
when demoting, which may cause the cpuset settings to violated.
Implement cpuset_node_allowed() to check the cpuset.mems_effective
associated wih the mem_cgroup of the lruvec being scanned. This only
applies to cgroup/cpuset v2, as cpuset exists in a different hierarchy
than mem_cgroup in v1.
This requires renaming the existing cpuset_node_allowed() to be
cpuset_current_now_allowed() - which is more descriptive anyway - to
implement the new cpuset_node_allowed() which takes a target cgroup.
This patch (of 2):
Rename cpuset_node_allowed to reflect that the function checks the current
task's cpuset.mems. This allows us to make a new cpuset_node_allowed
function that checks a target cgroup's cpuset.mems.
Link: https://lkml.kernel.org/r/20250424202806.52632-1-gourry@gourry.net
Link: https://lkml.kernel.org/r/20250424202806.52632-2-gourry@gourry.net
Signed-off-by: Gregory Price <gourry@gourry.net>
Acked-by: Waiman Long <longman@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michal Koutný <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Before introduction of ROX cache execmem allocation size was always
implicitly aligned to PAGE_SIZE inside vmalloc.
However, when allocation happens from the ROX cache, this is not
enforced.
Make sure that the allocation size is always consistently aligned to
PAGE_SIZE.
Mike said:
: Right now it'll make the maple trees in execmem_cache more compact.
: And it's a precaution for the case when execmem callers would want to
: change permissions on unaligned range because that would WARN_ON()
: loudly.
Peter said
: It should not have a runtime effect -- currently all this code is used
: with PAGE_SIZE multiples and everything just works. But whilst I was
: perusing this code, I noticed that nothing actually enforced this. If
: someone were to break this assumption things will go sideways.
Link: https://lkml.kernel.org/r/20250423144808.1619863-1-rppt@kernel.org
Fixes: 2e45474ab14f ("execmem: add support for cache of large ROX pages")
Signed-off-by: Mike Rapoport (Microsoft) <rppt@kernel.org>
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
In codes of alloc_vmap_area(), it returns the upper bound 'vend' to
indicate if the allocation is successful or failed. That is not very
clear.
Here change to return explicit error values and check them to judge if
allocation is successful.
Link: https://lkml.kernel.org/r/20250418223653.243436-6-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Shivank Garg <shivankg@amd.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Remove unneeded local variables and replace them with values.
Link: https://lkml.kernel.org/r/20250418223653.243436-5-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Reviewed-by: Shivank Garg <shivankg@amd.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When purge lazily freed vmap areas, VA stored in vn->pool[] will also be
taken away into free vmap tree partially or completely accordingly, that
is done in decay_va_pool_node(). When doing that, for each pool of node,
the whole list is detached from the pool for handling. At this time, that
pool is empty. It's not necessary to update the pool size each time when
one VA is removed and addded into free vmap tree.
Here change code to update the pool size when attaching the pool back.
Link: https://lkml.kernel.org/r/20250418223653.243436-4-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Shivank Garg <shivankg@amd.com>
Cc: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
When finding VA in vn->busy, if VA spans several zones and the passed addr
is not the same as va->va_start, we should scan the vn in reverse odrdr
because the starting address of VA must be smaller than the passed addr if
it really resides in the VA.
E.g on a system nr_vmap_nodes=100,
<----va---->
-|-----|-----|-----|-----|-----|-----|-----|-----|-----|-
... n-1 n n+1 n+2 ... 100 0 1
VA resides in node 'n' whereas it spans 'n', 'n+1' and 'n+2'. If passed
addr is within 'n+2', we should try nodes backwards on 'n+1' and 'n', then
succeed very soon.
Meanwhile we still need loop around because VA could spans node from 'n'
to node 100, node 0, node 1.
Anyway, changing to find in reverse order can improve efficiency on many
CPUs system.
Link: https://lkml.kernel.org/r/20250418223653.243436-3-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Shivank Garg <shivankg@amd.com>
Cc: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "mm/vmalloc.c: code cleanup and improvements", v2.
These changes were made from code inspection in mm/vmalloc.c.
This patch (of 5):
Static variable 'purge_ndoes' is defined in global scope, while it's only
used in function __purge_vmap_area_lazy(). It mainly serves to avoid
memory allocation repeatedly, especially when NR_CPUS is big.
While a local static variable can also satisfy the demand, and can improve
code readibility. Hence move its definition into
__purge_vmap_area_lazy().
Link: https://lkml.kernel.org/r/20250418223653.243436-1-bhe@redhat.com
Link: https://lkml.kernel.org/r/20250418223653.243436-2-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Reviewed-by: Shivank Garg <shivankg@amd.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Use cl@gentwo.org throughout and remove the old email addresses.
Link: https://lkml.kernel.org/r/8b962f57-4d98-cbb0-cd82-b6ba456733e8@gentwo.org
Signed-off-by: Christoph Lameter <cl@gentwo.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Corrected minor typos in comments:
- 'contigious' -> 'contiguous'
- 'hierarcy' -> 'hierarchy'
This is a non-functional change in comment text only.
Link: https://lkml.kernel.org/r/20250420140440.18817-1-pvkumar5749404@gmail.com
Signed-off-by: Prabhav Kumar Vaish <pvkumar5749404@gmail.com>
Reviewed-by: Mike Rapoport (Microsoft) <rppt@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
DAMON sysfs interface file for DAMOS quota goal's node id argument is not
passed to core layer. Implement the link.
Link: https://lkml.kernel.org/r/20250420194030.75838-4-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Yunjeong Mun <yunjeong.mun@sk.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
DAMOS_QUOTA_NODE_MEM_{USED,FREE}_BP DAMOS quota goal metrics require the
node id parameter. However, there is no DAMON user ABI for setting it.
Implement a DAMON sysfs file for that with name 'nid', under the quota
goal directory.
Link: https://lkml.kernel.org/r/20250420194030.75838-3-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Yunjeong Mun <yunjeong.mun@sk.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "mm/damon: auto-tune DAMOS for NUMA setups including tiered
memory".
Utilizing DAMON for memory tiering usually requires manual tuning and/or
tedious controls. Let it self-tune hotness and coldness thresholds for
promotion and demotion aiming high utilization of high memory tiers, by
introducing new DAMOS quota goal metrics representing the used and the
free memory ratios of specific NUMA nodes. And introduce a sample DAMON
module that demonstrates how the new feature can be used for memory
tiering use cases.
Backgrounds
===========
A type of tiered memory system exposes the memory tiers as NUMA nodes. A
straightforward pages placement strategy for such systems is placing
access-hot and cold pages on upper and lower tiers, reespectively,
pursuing higher utilization of upper tiers. Since access temperature can
be dynamic, periodically finding and migrating hot pages and cold pages to
proper tiers (promoting and demoting) is also required. Linux kernel
provides several features for such dynamic and transparent pages
placement.
Page Faults and LRU
-------------------
One widely known way is using NUMA balancing in tiering mode (a.k.a
NUMAB-2) and reclaim-based demotion features. In the setup, NUMAB-2 finds
hot pages using access check-purpose page faults (a.k.a prot_none) and
promote those inside each process' context, until there is no more pages
to promote, or the upper tier is filled up and memory pressure happens.
In the latter case, LRU-based reclaim logic wakes up as a response to the
memory pressure and demotes cold pages to lower tiers in asynchronous
(kswapd) and/or synchronous ways (direct reclaim).
DAMON
-----
Yet another available solution is using DAMOS with migrate_hot and
migrate_cold DAMOS actions for promotions and demotions, respectively. To
make it optimum, users need to specify aggressiveness and access
temperature thresholds for promotions and demotions in a good balance that
results in high utilization of upper tiers. The number of parameters is
not small, and optimum parameter values depend on characteristics of the
underlying hardware and the workload. As a result, it often requires
manual, time consuming and repetitive tuning of the DAMOS schemes for
given workloads and systems combinations.
Self-tuned DAMON-based Memory Tiering
=====================================
To solve such manual tuning problems, DAMOS provides aim-oriented
feedback-driven quotas self-tuning. Using the feature, we design a
self-tuned DAMON-based memory tiering for general multi-tier memory
systems.
For each memory tier node, if it has a lower tier, run a DAMOS scheme that
demotes cold pages of the node, auto-tuning the aggressiveness aiming an
amount of free space of the node. The free space is for keeping the
headroom that avoids significant memory pressure during upper tier memory
usage spike, and promoting hot pages from the lower tier.
For each memory tier node, if it has an upper tier, run a DAMOS scheme
that promotes hot pages of the current node to the upper tier, auto-tuning
the aggressiveness aiming a high utilization ratio of the upper tier. The
target ratio is to ensure higher tiers are utilized as much as possible.
It should match with the headroom for demotion scheme, but have slight
overlap, to ensure promotion and demotion are not entirely stopped.
The aim-oriented aggressiveness auto-tuning of DAMOS is already available.
Hence, to make such tiering solution implementation, only new quota goal
metrics for utilization and free space ratio of specific NUMA node need to
be developed.
Discussions
===========
The design imposes below discussion points.
Expected Behaviors
------------------
The system will let upper tier memory node accommodates as many hot data
as possible. If total amount of the data is less than the top tier
memory's promotion/demotion target utilization, entire data will be just
placed on the top tier. Promotion scheme will do nothing since there is
no data to promote. Demotion scheme will also do nothing since the free
space ratio of the top tier is higher than the goal.
Only if the amount of data is larger than the top tier's utilization
ratio, demotion scheme will demote cold pages and ensure the headroom free
space. Since the promotion and demotion schemes for a single node has
small overlap at their target utilization and free space goals, promotions
and demotions will continue working with a moderate aggressiveness level.
It will keep all data is placed on access hotness under dynamic access
pattern, while minimizing the migration overhead.
In any case, each node will keep headroom free space and as many upper
tiers are utilized as possible.
Ease of Use
-----------
Users still need to set the target utilization and free space ratio, but
it will be easier to set. We argue 99.7 % utilization and 0.5 % free
space ratios can be good default values. It can be easily adjusted based
on desired headroom size of given use case. Users are also still required
to answer the minimum coldness and hotness thresholds. Together with
monitoring intervals auto-tuning[2], DAMON will always show meaningful
amount of hot and cold memory. And DAMOS quota's prioritization mechanism
will make good decision as long as the source information is that
colorful. Hence, users can very naively set the minimum criterias. We
believe any access observation and no access observation within last one
aggregation interval is enough for minimum hot and cold regions criterias.
General Tiered Memory Setup Applicability
-----------------------------------------
The design can be applied to any number of tiers having any performance
characteristics, as long as they can be hierarchical. Hence, applying the
system to different tiered memory system will be straightforward. Note
that this assumes only single CPU NUMA node case. Because today's DAMON
is not aware of which CPU made each access, applying this on systems
having multiple CPU NUMA nodes can be complicated. We are planning to
extend DAMON for the use case, but that's out of the scope of this patch
series.
How To Use
----------
Users can implement the auto-tuned DAMON-based memory tiering using DAMON
sysfs interface. It can be easily done using DAMON user-space tool like
user-space tool. Below evaluation results section shows an example DAMON
user-space tool command for that.
For wider and simpler deployment, having a kernel module that sets up and
run the DAMOS schemes via DAMON kernel API can be useful. The module can
enable the memory tiering at boot time via kernel command line parameter
or at run time with single command. This patch series implements a sample
DAMON kernel module that shows how such module can be implemented.
Comparison To Page Faults and LRU-based Approaches
--------------------------------------------------
The existing page faults based promotion (NUMAB-2) does hot pages
detection and migration in the process context. When there are many pages
to promote, it can block the progress of the application's real works.
DAMOS works in asynchronous worker thread, so it doesn't block the real
works.
NUMAB-2 doesn't provide a way to control aggressiveness of promotion other
than the maximum amount of pages to promote per given time widnow. If hot
pages are found, promotions can happen in the upper-bound speed,
regardless of upper tier's memory pressure. If the maximum speed is not
well set for the given workload, it can result in slow promotion or
unnecessary memory pressure. Self-tuned DAMON-based memory tiering
alleviates the problem by adjusting the speed based on current utilization
of the upper tier.
LRU-based demotion can be triggered in both asynchronous (kswapd) and
synchronous (direct reclaim) ways. Other than the way of finding cold
pages, asynchronous LRU-based demotion and DAMON-based demotion has no big
difference. DAMON-based demotion can make a better balancing with
DAMON-based promotion, though. The LRU-based demotion can do better than
DAMON-based demotion when the tier is having significant memory pressure.
It would be wise to use DAMON-based demotion as a proactive and primary
one, but utilizing LRU-based demotions together as a fast backup solution.
Evaluation
==========
In short, under a setup that requires fast and frequent promotions,
self-tuned DAMON-based memory tiering's hot pages promotion improves
performance about 4.42 %. We believe this shows self-tuned DAMON-based
promotion's effectiveness. Meanwhile, NUMAB-2's hot pages promotion
degrades the performance about 7.34 %. We suspect the degradation is
mostly due to NUMAB-2's synchronous nature that can block the
application's progress, which highlights the advantage of DAMON-based
solution's asynchronous nature.
Note that the test was done with the RFC version of this patch series. We
don't run it again since this patch series got no meaningful change after
the RFC, while the test takes pretty long time.
Setup
-----
Hardware. Use a machine that equips 250 GiB DRAM memory tier and 50 GiB
CXL memory tier. The tiers are exposed as NUMA nodes 0 and 1,
respectively.
Kernel. Use Linux kernel v6.13 that modified as following. Add all DAMON
patches that available on mm tree of 2025-03-15, and this patch series.
Also modify it to ignore mempolicy() system calls, to avoid bad effects
from application's traditional NUMA systems assumed optimizations.
Workload. Use a modified version of Taobench benchmark[3] that available
on DCPerf benchmark suite. It represents an in-memory caching workload.
We set its 'memsize', 'warmup_time', and 'test_time' parameter as 340 GiB,
2,500 seconds and 1,440 seconds. The parameters are chosen to ensure the
workload uses more than DRAM memory tier. Its RSS under the parameter
grows to 270 GiB within the warmup time.
It turned out the workload has a very static access pattrn. Only about 13
% of the RSS is frequently accessed from the beginning to end. Hence
promotion shows no meaningful performance difference regardless of
different design and implementations. We therefore modify the kernel to
periodically demote up to 10 GiB hot pages and promote up to 10 GiB cold
pages once per minute. The intention is to simulate periodic access
pattern changes. The hotness and coldness threshold is very naively set
so that it is more like random access pattern change rather than strict
hot/cold pages exchange. This is why we call the workload as "modified".
It is implemented as two DAMOS schemes each running on an asynchronous
thread. It can be reproduced with DAMON user-space tool like below.
# ./damo start \
--ops paddr --numa_node 0 --monitoring_intervals 10s 200s 200s \
--damos_action migrate_hot 1 \
--damos_quota_interval 60s --damos_quota_space 10G \
--ops paddr --numa_node 1 --monitoring_intervals 10s 200s 200s \
--damos_action migrate_cold 0 \
--damos_quota_interval 60s --damos_quota_space 10G \
--nr_schemes 1 1 --nr_targets 1 1 --nr_ctxs 1 1
System configurations. Use below variant system configurations.
- Baseline. No memory tiering features are turned on.
- Numab_tiering. On the baseline, enable NUMAB-2 and relcaim-based
demotion. In detail, following command is executed:
echo 2 > /proc/sys/kernel/numa_balancing;
echo 1 > /sys/kernel/mm/numa/demotion_enabled;
echo 7 > /proc/sys/vm/zone_reclaim_mode
- DAMON_tiering. On the baseline, utilize self-tuned DAMON-based memory
tiering implementation via DAMON user-space tool. It utilizes two
kernel threads, namely promotion thread and demotion thread. Demotion
thread monitors access pattern of DRAM node using DAMON with
auto-tuned monitoring intervals aiming 4% DAMON-observed access ratio,
and demote coldest pages up to 200 MiB per second aiming 0.5% free
space of DRAM node. Promotion thread monitors CXL node using same
intervals auto-tuning, and promote hot pages in same way but aiming
for 99.7% utilization of DRAM node. Because DAMON provides only
best-effort accuracy, add young page DAMOS filters to allow only and
reject all young pages at promoting and demoting, respectively. It
can be reproduced with DAMON user-space tool like below.
# ./damo start \
--numa_node 0 --monitoring_intervals_goal 4% 3 5ms 10s \
--damos_action migrate_cold 1 --damos_access_rate 0% 0% \
--damos_apply_interval 1s \
--damos_quota_interval 1s --damos_quota_space 200MB \
--damos_quota_goal node_mem_free_bp 0.5% 0 \
--damos_filter reject young \
--numa_node 1 --monitoring_intervals_goal 4% 3 5ms 10s \
--damos_action migrate_hot 0 --damos_access_rate 5% max \
--damos_apply_interval 1s \
--damos_quota_interval 1s --damos_quota_space 200MB \
--damos_quota_goal node_mem_used_bp 99.7% 0 \
--damos_filter allow young \
--damos_nr_quota_goals 1 1 --damos_nr_filters 1 1 \
--nr_targets 1 1 --nr_schemes 1 1 --nr_ctxs 1 1
Measurment Results
------------------
On each system configuration, run the modified version of Taobench and
collect 'score'. 'score' is a metric that calculated and provided by
Taobench to represents the performance of the run on the system. To
handle the measurement errors, repeat the measurement five times. The
results are as below.
Config Score Stdev (%) Normalized
Baseline 1.6165 0.0319 1.9764 1.0000
Numab_tiering 1.4976 0.0452 3.0209 0.9264
DAMON_tiering 1.6881 0.0249 1.4767 1.0443
'Config' column shows the system config of the measurement. 'Score'
column shows the 'score' measurement in average of the five runs on the
system config. 'Stdev' column shows the standsard deviation of the five
measurements of the scores. '(%)' column shows the 'Stdev' to 'Score'
ratio in percentage. Finally, 'Normalized' column shows the averaged
score values of the configs that normalized to that of 'Baseline'.
The periodic hot pages demotion and cold pages promotion that was
conducted to simulate dynamic access pattern was started from the
beginning of the workload. It resulted in the DRAM tier utilization
always under the watermark, and hence no real demotion was happened for
all test runs. This means the above results show no difference between
LRU-based and DAMON-based demotions. Only difference between NUMAB-2 and
DAMON-based promotions are represented on the results.
Numab_tiering config degraded the performance about 7.36 %. We suspect
this happened because NUMAB-2's synchronous promotion was blocking the
Taobench's real work progress.
DAMON_tiering config improved the performance about 4.43 %. We believe
this shows effectiveness of DAMON-based promotion that didn't block
Taobench's real work progress due to its asynchronous nature. Also this
means DAMON's monitoring results are accurate enough to provide visible
amount of improvement.
Evaluation Limitations
----------------------
As mentioned above, this evaluation shows only comparison of promotion
mechanisms. DAMON-based tiering is recommended to be used together with
reclaim-based demotion as a faster backup under significant memory
pressure, though.
From some perspective, the modified version of Taobench may seems making
the picture distorted too much. It would be better to evaluate with more
realistic workload, or more finely tuned micro benchmarks.
Patch Sequence
==============
The first patch (patch 1) implements two new quota goal metrics on core
layer and expose it to DAMON core kernel API. The second and third ones
(patches 2 and 3) further link it to DAMON sysfs interface. Three
following patches (patches 4-6) document the new feature and sysfs file on
design, usage, and ABI documents. The final one (patch 7) implements a
working version of a self-tuned DAMON-based memory tiering solution in an
incomplete but easy to understand form as a kernel module under
samples/damon/ directory.
References
==========
[1] https://lore.kernel.org/20231112195602.61525-1-sj@kernel.org/
[2] https://lore.kernel.org/20250303221726.484227-1-sj@kernel.org
[3] https://github.com/facebookresearch/DCPerf/blob/main/packages/tao_bench/README.md
This patch (of 7):
Used and free space ratios for specific NUMA nodes can be useful inputs
for NUMA-specific DAMOS schemes' aggressiveness self-tuning feedback loop.
Implement DAMOS quota goal metrics for such self-tuned schemes.
Link: https://lkml.kernel.org/r/20250420194030.75838-1-sj@kernel.org
Link: https://lkml.kernel.org/r/20250420194030.75838-2-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Cc: Yunjeong Mun <yunjeong.mun@sk.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull vfs fixes from Christian Brauner:
- Ensure that simple_xattr_list() always includes security.* xattrs
- Fix eventpoll busy loop optimization when combined with timeouts
- Disable swapon() for devices with block sizes greater than page sizes
- Don't call errseq_set() twice during mark_buffer_write_io_error().
Just use mapping_set_error() which takes care to not deference
unconditionally
* tag 'vfs-6.15-rc7.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
fs: Remove redundant errseq_set call in mark_buffer_write_io_error.
swapfile: disable swapon for bs > ps devices
fs/eventpoll: fix endless busy loop after timeout has expired
fs/xattr.c: fix simple_xattr_list to always include security.* xattrs
|
|
HMM callers use PFN list to populate range while calling
to hmm_range_fault(), the conversion from PFN to DMA address
is done by the callers with help of another DMA list. However,
it is wasteful on any modern platform and by doing the right
logic, that DMA list can be avoided.
Provide generic logic to manage these lists and gave an interface
to map/unmap PFNs to DMA addresses, without requiring from the callers
to be an experts in DMA core API.
Tested-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
|
|
Introduce new sticky flag (HMM_PFN_DMA_MAPPED), which isn't overwritten
by HMM range fault. Such flag allows users to tag specific PFNs with
information if this specific PFN was already DMA mapped.
Tested-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
|
|
strncpy_from_user_nofault() should return the length of the copied string
including the trailing NUL, but if the argument unsafe_addr points to an
empty string ({'\0'}), the return value is 0.
This happens as strncpy_from_user() copies terminal symbol into dst and
returns 0 (as expected), but strncpy_from_user_nofault does not modify ret
as it is not equal to count and not greater than 0, so 0 is returned,
which contradicts the contract.
Link: https://lkml.kernel.org/r/20250422131449.57177-1-mykyta.yatsenko5@gmail.com
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Reviewed-by: Andrii Nakryiko <andrii@kernel.org>
Cc: "Masami Hiramatsu (Google)" <mhiramat@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The weighted interleave policy distributes page allocations across
multiple NUMA nodes based on their performance weight, thereby improving
memory bandwidth utilization. The weight values for each node are
configured through sysfs.
Previously, sysfs entries for configuring weighted interleave were created
for all possible nodes (N_POSSIBLE) at initialization, including nodes
that might not have memory. However, not all nodes in N_POSSIBLE are
usable at runtime, as some may remain memoryless or offline. This led to
sysfs entries being created for unusable nodes, causing potential
misconfiguration issues.
To address this issue, this patch modifies the sysfs creation logic to:
1) Limit sysfs entries to nodes that are online and have memory, avoiding
the creation of sysfs entries for nodes that cannot be used.
2) Support memory hotplug by dynamically adding and removing sysfs entries
based on whether a node transitions into or out of the N_MEMORY state.
Additionally, the patch ensures that sysfs attributes are properly managed
when nodes go offline, preventing stale or redundant entries from
persisting in the system.
By making these changes, the weighted interleave policy now manages its
sysfs entries more efficiently, ensuring that only relevant nodes are
considered for interleaving, and dynamically adapting to memory hotplug
events.
[dan.carpenter@linaro.org: fix error code in sysfs_wi_node_add()]
Link: https://lkml.kernel.org/r/aBjL7Bwc0QBzgajK@stanley.mountain
Link: https://lkml.kernel.org/r/20250417072839.711-4-rakie.kim@sk.com
Co-developed-by: Honggyu Kim <honggyu.kim@sk.com>
Signed-off-by: Honggyu Kim <honggyu.kim@sk.com>
Co-developed-by: Yunjeong Mun <yunjeong.mun@sk.com>
Signed-off-by: Yunjeong Mun <yunjeong.mun@sk.com>
Signed-off-by: Rakie Kim <rakie.kim@sk.com>
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Joshua Hahn <joshua.hahnjy@gmail.com>
Reviewed-by: Gregory Price <gourry@gourry.net>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: "Huang, Ying" <ying.huang@linux.alibaba.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Previously, the weighted interleave sysfs structure was statically managed
during initialization. This prevented new nodes from being recognized
when memory hotplug events occurred, limiting the ability to update or
extend sysfs entries dynamically at runtime.
To address this, this patch refactors the sysfs infrastructure and
encapsulates it within a new structure, `sysfs_wi_group`, which holds both
the kobject and an array of node attribute pointers.
By allocating this group structure globally, the per-node sysfs attributes
can be managed beyond initialization time, enabling external modules to
insert or remove node entries in response to events such as memory hotplug
or node online/offline transitions.
Instead of allocating all per-node sysfs attributes at once, the
initialization path now uses the existing sysfs_wi_node_add() and
sysfs_wi_node_delete() helpers. This refactoring makes it possible to
modularly manage per-node sysfs entries and ensures the infrastructure is
ready for runtime extension.
Link: https://lkml.kernel.org/r/20250417072839.711-3-rakie.kim@sk.com
Signed-off-by: Rakie Kim <rakie.kim@sk.com>
Reviewed-by: Gregory Price <gourry@gourry.net>
Reviewed-by: Joshua Hahn <joshua.hahnjy@gmail.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Honggyu Kim <honggyu.kim@sk.com>
Cc: "Huang, Ying" <ying.huang@linux.alibaba.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Yunjeong Mun <yunjeong.mun@sk.com>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "Enhance sysfs handling for memory hotplug in weighted
interleave", v9.
The following patch series enhances the weighted interleave policy in the
memory management subsystem by improving sysfs handling, fixing memory
leaks, and introducing dynamic sysfs updates for memory hotplug support.
This patch (of 3):
Memory leaks occurred when removing sysfs attributes for weighted
interleave. Improper kobject deallocation led to unreleased memory when
initialization failed or when nodes were removed.
The risk of leak is low because it only appears to trigger if setup
fails. Setup only fails due to -ENOMEM which is unlikely to happen
from a late_initcall() when memory pressure is low.
This patch resolves the issue by replacing unnecessary `kfree()` calls
with proper `kobject_del()` and `kobject_put()` sequences, ensuring
correct teardown and preventing memory leaks.
By explicitly calling `kobject_del()` before `kobject_put()`, the release
function is now invoked safely, and internal sysfs state is correctly
cleaned up. This guarantees that the memory associated with the kobject
is fully released and avoids resource leaks, thereby improving system
stability.
Additionally, sysfs_remove_file() is no longer called from the release
function to avoid accessing invalid sysfs state after kobject_del(). All
attribute removals are now done before kobject_del(), preventing WARN_ON()
in kernfs and ensuring safe and consistent cleanup of sysfs entries.
Link: https://lkml.kernel.org/r/20250417072839.711-1-rakie.kim@sk.com
Link: https://lkml.kernel.org/r/20250417072839.711-2-rakie.kim@sk.com
Fixes: dce41f5ae253 ("mm/mempolicy: implement the sysfs-based weighted_interleave interface")
Signed-off-by: Rakie Kim <rakie.kim@sk.com>
Reviewed-by: Gregory Price <gourry@gourry.net>
Reviewed-by: Joshua Hahn <joshua.hahnjy@gmail.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Dan Williams <dan.j.williams@intel.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Honggyu Kim <honggyu.kim@sk.com>
Cc: "Huang, Ying" <ying.huang@linux.alibaba.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Yunjeong Mun <yunjeong.mun@sk.com>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
free_percpu() checks for NULL pointers internally. Remove unneeded NULL
check here.
Link: https://lkml.kernel.org/r/20250417084330.937380-1-nichen@iscas.ac.cn
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently both atomics share one cache-line:
<snip>
...
ffffffff83eab400 b vmap_lazy_nr
ffffffff83eab408 b nr_vmalloc_pages
...
<snip>
those are global variables and they are only 8 bytes apart. Since they
are modified by different threads this causes a false sharing. This can
lead to a performance drop due to unnecessary cache invalidations.
After this patch it is aligned to a cache line boundary:
<snip>
...
ffffffff8260a600 d vmap_lazy_nr
ffffffff8260a640 d nr_vmalloc_pages
...
<snip>
Link: https://lkml.kernel.org/r/20250417161216.88318-4-urezki@gmail.com
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Adrian Huang <ahuang12@lenovo.com>
Tested-by: Adrian Huang <ahuang12@lenovo.com>
Cc: Mateusz Guzik <mjguzik@gmail.com>
Cc: Christop Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
After the check for queue_folio_required(), the code only cares about the
folio in the for loop, i.e the PTEs are redundant. Therefore, optimize
this loop by skipping over a PTE batch mapping the same folio.
With a test program migrating pages of the calling process, which includes
a mapped VMA of size 4GB with pte-mapped large folios of order-9, and
migrating once back and forth node-0 and node-1, the average execution
time reduces from 7.5 to 4 seconds, giving an approx 47% speedup.
Link: https://lkml.kernel.org/r/20250416053048.96479-1-dev.jain@arm.com
Signed-off-by: Dev Jain <dev.jain@arm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Cc: Yang Shi <yang@os.amperecomputing.com>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently the VMA and mmap locking logic is entangled in two of the most
overwrought files in mm - include/linux/mm.h and mm/memory.c. Separate
this logic out so we can more easily make changes and create an
appropriate MAINTAINERS entry that spans only the logic relating to
locking.
This should have no functional change. Care is taken to avoid dependency
loops, we must regrettably keep release_fault_lock() and
assert_fault_locked() in mm.h as a result due to the dependence on the
vm_fault type.
Additionally we must declare rcuwait_wake_up() manually to avoid a
dependency cycle on linux/rcuwait.h.
Additionally move the nommu implementatino of lock_mm_and_find_vma() to
mmap_lock.c so everything lock-related is in one place.
Link: https://lkml.kernel.org/r/bec6c8e29fa8de9267a811a10b1bdae355d67ed4.1744799282.git.lorenzo.stoakes@oracle.com
Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Reviewed-by: Suren Baghdasaryan <surenb@google.com>
Reviewed-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: "Paul E . McKenney" <paulmck@kernel.org>
Cc: SeongJae Park <sj@kernel.org>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Memory cgroup accounting is expensive and to reduce the cost, the kernel
maintains per-cpu charge cache for a single memcg. So, if a charge
request comes for a different memcg, the kernel will flush the old memcg's
charge cache and then charge the newer memcg a fixed amount (64 pages),
subtracts the charge request amount and stores the remaining in the
per-cpu charge cache for the newer memcg.
This mechanism is based on the assumption that the kernel, for locality,
keep a process on a CPU for long period of time and most of the charge
requests from that process will be served by that CPU's local charge
cache.
However this assumption breaks down for incoming network traffic in a
multi-tenant machine. We are in the process of running multiple workloads
on a single machine and if such workloads are network heavy, we are seeing
very high network memory accounting cost. We have observed multiple CPUs
spending almost 100% of their time in net_rx_action and almost all of that
time is spent in memcg accounting of the network traffic.
More precisely, net_rx_action is serving packets from multiple workloads
and is observing/serving mix of packets of these workloads. The memcg
switch of per-cpu cache is very expensive and we are observing a lot of
memcg switches on the machine. Almost all the time is being spent on
charging new memcg and flushing older memcg cache. So, definitely we need
per-cpu cache that support multiple memcgs for this scenario.
This patch implements a simple (and dumb) multiple memcg percpu charge
cache. Actually we started with more sophisticated LRU based approach but
the dumb one was always better than the sophisticated one by 1% to 3%, so
going with the simple approach.
Some of the design choices are:
1. Fit all caches memcgs in a single cacheline.
2. The cache array can be mix of empty slots or memcg charged slots, so
the kernel has to traverse the full array.
3. The cache drain from the reclaim will drain all cached memcgs to keep
things simple.
To evaluate the impact of this optimization, on a 72 CPUs machine, we ran
the following workload where each netperf client runs in a different
cgroup. The next-20250415 kernel is used as base.
$ netserver -6
$ netperf -6 -H ::1 -l 60 -t TCP_SENDFILE -- -m 10K
number of clients | Without patch | With patch
6 | 42584.1 Mbps | 48603.4 Mbps (14.13% improvement)
12 | 30617.1 Mbps | 47919.7 Mbps (56.51% improvement)
18 | 25305.2 Mbps | 45497.3 Mbps (79.79% improvement)
24 | 20104.1 Mbps | 37907.7 Mbps (88.55% improvement)
30 | 14702.4 Mbps | 30746.5 Mbps (109.12% improvement)
36 | 10801.5 Mbps | 26476.3 Mbps (145.11% improvement)
The results show drastic improvement for network intensive workloads.
[shakeel.butt@linux.dev: add BUILD_BUG_ON() for MEMCG_CHARGE_BATCH]
Link: https://lkml.kernel.org/r/rlsgeosg3j7v5nihhbxxxbv3xfy4ejvigihj7lkkbt3n6imyne@2apxx2jm2e57
[shakeel.butt@linux.dev: simplify refill_stock]
Link: https://lkml.kernel.org/r/as5cdsm4lraxupg3t6onep2ixql72za25hvd4x334dsoyo4apr@zyzl4vkuevuv
[hughd@google.com: it's better to stock nr_pages than the uninitialized stock_pages]
Link: https://lkml.kernel.org/r/d542d18f-1caa-6fea-e2c3-3555c87bcf64@google.com
[shakeel.butt@linux.dev: add comment per Michal and use DEFINE_PER_CPU_ALIGNED instead of DEFINE_PER_CPU per Vlastimil]
Link: https://lkml.kernel.org/r/dieeei3squ2gcnqxdjayvxbvzldr266rhnvtl3vjzsqevxkevf@ckui5vjzl2qg
Link: https://lkml.kernel.org/r/20250416180229.2902751-1-shakeel.butt@linux.dev
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Eric Dumaze <edumazet@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Soheil Hassas Yeganeh <soheil@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
free_page_and_swap_cache() takes a struct page pointer as input parameter,
but it will immediately convert it to folio and all operations following
within use folio instead of page. It makes more sense to pass in folio
directly.
Convert free_page_and_swap_cache() to free_folio_and_swap_cache() to
consume folio directly.
Link: https://lkml.kernel.org/r/20250416201720.41678-1-nifan.cxl@gmail.com
Signed-off-by: Fan Ni <fan.ni@samsung.com>
Acked-by: Davidlohr Bueso <dave@stgolabs.net>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Adam Manzanares <a.manzanares@samsung.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@kernel.org>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Luis Chamberalin <mcgrof@kernel.org>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
commit c928807f6f6b6("mm/page_alloc: keep track of free highatomic")
adds a new variable nr_free_highatomic, which is useful for analyzing low
mem issues. add nr_free_highatomic in show_free_areas.
Signed-off-by: gao xu <gaoxu2@honor.com>
Link: https://lkml.kernel.org/r/d92eeff74f7a4578a14ac777cfe3603a@honor.com
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Barry Song <baohua@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Yu Zhao <yuzhao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The scan and total_scan variables can be initialized to 0 when they are
defined, replacing the separate assignment statements.
Link: https://lkml.kernel.org/r/20250417092422.1333620-1-hao.ge@linux.dev
Signed-off-by: Hao Ge <gehao@kylinos.cn>
Acked-by: Dev Jain <dev.jain@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The code style in fault_in_readable() and fault_in_writable() is a little
inconsistent with fault_in_safe_writeable(). In fault_in_readable() and
fault_in_writable(), it uses 'uaddr' passed in as loop cursor. While in
fault_in_safe_writeable(), local variable 'start' is used as loop cursor.
This may mislead people when reading code or making change in these codes.
Here define explicit loop cursor and use for loop to simplify codes in
these three functions. These cleanup can make them be consistent in code
style and improve readability.
[bhe@redhat.com: address minor concerns from David]
Link: https://lkml.kernel.org/r/Z/sbv3EmLXWgEE7+@MiWiFi-R3L-srv
Link: https://lkml.kernel.org/r/20250410035717.473207-5-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Yanjun.Zhu <yanjun.zhu@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
In the current kernel, only pud huge page is supported in some
architectures. P4d and pgd huge pages haven't been supported yet. And in
mm/gup.c, there's no pgd huge page handling in the follow_page_mask() code
path. Hence it doesn't make sense to only have gup_fast_pgd_leaf() in
gup_fast code path.
Here remove gup_fast_pgd_leaf() and clean up the relevant codes.
Link: https://lkml.kernel.org/r/20250410035717.473207-4-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Yanjun.Zhu <yanjun.zhu@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "mm/gup: Minor fix, cleanup and improvements", v4.
These were made from code inspection in mm/gup.c.
This patch (of 3):
In __get_user_pages(), it will traverse page table and take a reference to
the page the given user address corresponds to if GUP_GET or GUP_PIN is
set. However, it's not supported both GUP_GET and GUP_PIN are set. Even
though this check need be done, it should be done earlier, but not doing
it till entering into follow_page_pte() and failed.
Furthermore, this checking has been done in is_valid_gup_args() and all
external users of __get_user_pages() will call is_valid_gup_args() to
catch the illegal setting. We don't need to worry about internal users of
__get_user_pages() because the gup_flags are set by MM code correctly.
Here remove the checking in follow_page_pte(), and add VM_WARN_ON_ONCE()
to catch the possible exceptional setting just in case.
And also change the VM_BUG_ON to VM_WARN_ON_ONCE() for checking (!!pages
!= !!(gup_flags & (FOLL_GET | FOLL_PIN))) because the checking has been
done in is_valid_gup_args() for external users of __get_user_pages().
Link: https://lkml.kernel.org/r/20250410035717.473207-1-bhe@redhat.com
Link: https://lkml.kernel.org/r/20250410035717.473207-3-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Yanjun.Zhu <yanjun.zhu@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
alloc_buddy_hugetlb_folio() allocates a rmappable folio, then strips the
rmappable part and freezes it. We can simplify all that by allocating
frozen pages directly.
Link: https://lkml.kernel.org/r/20250411132359.312708-1-osalvador@suse.de
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Suggested-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Muchun Song <muchun.song@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Switch from the atomic_long_add_return() to its relaxed version.
We do not need a full memory barrier or any memory ordering during
increasing the "vmap_lazy_nr" variable. What we only need is to do it
atomically. This is what atomic_long_add_return_relaxed() guarantees.
AARCH64:
<snip>
Default:
40ec: d34cfe94 lsr x20, x20, #12
40f0: 14000044 b 4200 <free_vmap_area_noflush+0x19c>
40f4: 94000000 bl 0 <__sanitizer_cov_trace_pc>
40f8: 90000000 adrp x0, 0 <__traceiter_alloc_vmap_area>
40fc: 91000000 add x0, x0, #0x0
4100: f8f40016 ldaddal x20, x22, [x0]
4104: 8b160296 add x22, x20, x22
Relaxed:
40ec: d34cfe94 lsr x20, x20, #12
40f0: 14000044 b 4200 <free_vmap_area_noflush+0x19c>
40f4: 94000000 bl 0 <__sanitizer_cov_trace_pc>
40f8: 90000000 adrp x0, 0 <__traceiter_alloc_vmap_area>
40fc: 91000000 add x0, x0, #0x0
4100: f8340016 ldadd x20, x22, [x0]
4104: 8b160296 add x22, x20, x22
<snip>
Link: https://lkml.kernel.org/r/20250415112646.113091-1-urezki@gmail.com
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Reviewed-by: Baoquan He <bhe@redhat.com>
Cc: Christop Hellwig <hch@infradead.org>
Cc: Mateusz Guzik <mjguzik@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Before trying to allocate a page, gather_surplus_pages() sets up a
nodemask for the nodes we can allocate from, but instead of passing the
nodemask down the road to the page allocator, it iterates over the nodes
within that nodemask right there, meaning that the page allocator will
receive a preferred_nid and a null nodemask.
This is a problem when using a memory policy, because it might be that the
page allocator ends up using a node as a fallback which is not represented
in the policy.
Avoid that by passing the nodemask directly to the page allocator, so it
can filter out fallback nodes that are not part of the nodemask.
Link: https://lkml.kernel.org/r/20250415121503.376811-1-osalvador@suse.de
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: David Hildenbrand <david@redhat.com>
Cc: Muchun Song <muchun.song@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently the kernel maintains the stats updates per-memcg which is needed
to implement stats flushing threshold. On the update side, the update is
added to the per-cpu per-memcg update of the given memcg and all of its
ancestors. However when the given memcg has passed the flushing
threshold, all of its ancestors should have passed the threshold as well.
There is no need to traverse up the memcg tree to maintain the stats
updates.
Perf profile collected from our fleet shows that memcg_rstat_updated is
one of the most expensive memcg function i.e. a lot of cumulative CPU is
being spent on it. So, even small micro optimizations matter a lot. This
patch is microbenchmarked with multiple instances of netperf on a single
machine with locally running netserver and we see couple of percentage of
improvement.
Link: https://lkml.kernel.org/r/20250410025752.92159-1-shakeel.butt@linux.dev
Signed-off-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Roman Gushchin <roman.gushchin@linux.dev>
Reviewed-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
MADV_DONTNEED[_LOCKED] handling for [process_]madvise() flushes tlb for
each vma of each address range. Update the logic to do tlb flushes in a
batched way. Initialize an mmu_gather object from do_madvise() and
vector_madvise(), which are the entry level functions for
[process_]madvise(), respectively. And pass those objects to the function
for per-vma work, via madvise_behavior struct. Make the per-vma logic not
flushes tlb on their own but just saves the tlb entries to the received
mmu_gather object. For this internal logic change, make
zap_page_range_single_batched() non-static and use it directly from
madvise_dontneed_single_vma(). Finally, the entry level functions flush
the tlb entries that gathered for the entire user request, at once.
Link: https://lkml.kernel.org/r/20250410000022.1901-5-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Some of zap_page_range_single() callers such as [process_]madvise() with
MADV_DONTNEED[_LOCKED] cannot batch tlb flushes because
zap_page_range_single() flushes tlb for each invocation. Split out the
body of zap_page_range_single() except mmu_gather object initialization
and gathered tlb entries flushing for such batched tlb flushing usage.
To avoid hugetlb pages allocation failures from concurrent page faults,
the tlb flush should be done before hugetlb faults unlocking, though. Do
the flush and the unlock inside the split out function in the order for
hugetlb vma case. Refer to commit 2820b0f09be9 ("hugetlbfs: close race
between MADV_DONTNEED and page fault") for more details about the
concurrent faults' page allocation failure problem.
Link: https://lkml.kernel.org/r/20250410000022.1901-4-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
MADV_FREE handling for [process_]madvise() flushes tlb for each vma of
each address range. Update the logic to do tlb flushes in a batched way.
Initialize an mmu_gather object from do_madvise() and vector_madvise(),
which are the entry level functions for [process_]madvise(), respectively.
And pass those objects to the function for per-vma work, via
madvise_behavior struct. Make the per-vma logic not flushes tlb on their
own but just saves the tlb entries to the received mmu_gather object.
Finally, the entry level functions flush the tlb entries that gathered for
the entire user request, at once.
Link: https://lkml.kernel.org/r/20250410000022.1901-3-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Patch series "mm/madvise: batch tlb flushes for MADV_DONTNEED and
MADV_FREE", v3.
When process_madvise() is called to do MADV_DONTNEED[_LOCKED] or MADV_FREE
with multiple address ranges, tlb flushes happen for each of the given
address ranges. Because such tlb flushes are for the same process, doing
those in a batch is more efficient while still being safe. Modify
process_madvise() entry level code path to do such batched tlb flushes,
while the internal unmap logic do only gathering of the tlb entries to
flush.
In more detail, modify the entry functions to initialize an mmu_gather
object and pass it to the internal logic. And make the internal logic do
only gathering of the tlb entries to flush into the received mmu_gather
object. After all internal function calls are done, the entry functions
flush the gathered tlb entries at once.
Because process_madvise() and madvise() share the internal unmap logic,
make same change to madvise() entry code together, to make code consistent
and cleaner. It is only for keeping the code clean, and shouldn't degrade
madvise(). It could rather provide a potential tlb flushes reduction
benefit for a case that there are multiple vmas for the given address
range. It is only a side effect from an effort to keep code clean, so we
don't measure it separately.
Similar optimizations might be applicable to other madvise behavior such
as MADV_COLD and MADV_PAGEOUT. Those are simply out of the scope of this
patch series, though.
Patches Sequence
================
The first patch defines a new data structure for managing information that
is required for batched tlb flushes (mmu_gather and behavior), and update
code paths for MADV_DONTNEED[_LOCKED] and MADV_FREE handling internal
logic to receive it.
The second patch batches tlb flushes for MADV_FREE handling for both
madvise() and process_madvise().
Remaining two patches are for MADV_DONTNEED[_LOCKED] tlb flushes batching.
The third patch splits zap_page_range_single() for batching of
MADV_DONTNEED[_LOCKED] handling. The fourth patch batches tlb flushes for
the hint using the sub-logic that the third patch split out, and the
helpers for batched tlb flushes that introduced for the MADV_FREE case, by
the second patch.
Test Results
============
I measured the latency to apply MADV_DONTNEED advice to 256 MiB memory
using multiple process_madvise() calls. I apply the advice in 4 KiB sized
regions granularity, but with varying batch size per process_madvise()
call (vlen) from 1 to 1024. The source code for the measurement is
available at GitHub[1]. To reduce measurement errors, I did the
measurement five times.
The measurement results are as below. 'sz_batch' column shows the batch
size of process_madvise() calls. 'Before' and 'After' columns show the
average of latencies in nanoseconds that measured five times on kernels
that built without and with the tlb flushes batching of this series
(patches 3 and 4), respectively. For the baseline, mm-new tree of
2025-04-09[2] has been used, after reverting the second version of this
patch series and adding a temporal fix for !CONFIG_DEBUG_VM build
failure[3]. 'B-stdev' and 'A-stdev' columns show ratios of latency
measurements standard deviation to average in percent for 'Before' and
'After', respectively. 'Latency_reduction' shows the reduction of the
latency that the 'After' has achieved compared to 'Before', in percent.
Higher 'Latency_reduction' values mean more efficiency improvements.
sz_batch Before B-stdev After A-stdev Latency_reduction
1 146386348 2.78 111327360.6 3.13 23.95
2 108222130 1.54 72131173.6 2.39 33.35
4 93617846.8 2.76 51859294.4 2.50 44.61
8 80555150.4 2.38 44328790 1.58 44.97
16 77272777 1.62 37489433.2 1.16 51.48
32 76478465.2 2.75 33570506 3.48 56.10
64 75810266.6 1.15 27037652.6 1.61 64.34
128 73222748 3.86 25517629.4 3.30 65.15
256 72534970.8 2.31 25002180.4 0.94 65.53
512 71809392 5.12 24152285.4 2.41 66.37
1024 73281170.2 4.53 24183615 2.09 67.00
Unexpectedly the latency has reduced (improved) even with batch size one.
I think some of compiler optimizations have affected that, like also
observed with the first version of this patch series.
So, please focus on the proportion between the improvement and the batch
size. As expected, tlb flushes batching provides latency reduction that
proportional to the batch size. The efficiency gain ranges from about 33
percent with batch size 2, and up to 67 percent with batch size 1,024.
Please note that this is a very simple microbenchmark, so real efficiency
gain on real workload could be very different.
This patch (of 4):
To implement batched tlb flushes for MADV_DONTNEED[_LOCKED] and MADV_FREE,
an mmu_gather object in addition to the behavior integer need to be passed
to the internal logics. Using a struct can make it easy without
increasing the number of parameters of all code paths towards the internal
logic. Define a struct for the purpose and use it on the code path that
starts from madvise_do_behavior() and ends on madvise_dontneed_free().
Note that this changes madvise_walk_vmas() visitor type signature, too.
Specifically, it changes its 'arg' type from 'unsigned long' to the new
struct pointer.
Link: https://lkml.kernel.org/r/20250410000022.1901-1-sj@kernel.org
Link: https://lkml.kernel.org/r/20250410000022.1901-2-sj@kernel.org
Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: David Hildenbrand <david@redhat.com>
Cc: Liam R. Howlett <howlett@gmail.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: SeongJae Park <sj@kernel.org>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
