[RFC PATCH 0/5] add mTHP support for anonymous share pages (Baolin Wang) https://lkml.kernel.org/r/cover.1713755580.git.baolin.wang@linux.alibaba.com
Baolin adds mTHP support for anonymous shared pages with this RFC patchset. There are some todo items, though.
[LSF/MM/BPF TOPIC] SLAB BOF (Matthew Wilcox) https://lkml.kernel.org/r/ZiZy2VVpTQMjN-ko@casper.infradead.org
In this mail, Matthew introduces his radical idea around Slab allocator, which decouples the objects from the page and proposes a BoF to discuss it with not only Vlastimil but more audiences.
[RFC 0/3] Improve memory statistics for virtio balloon (zhenwei pi) https://lkml.kernel.org/r/20240415084113.1203428-1-pizhenwei@bytedance.com
This RFC patchset exposes six new VM statistics in the guest to the host via virtio-balloon. The statistics include oom-kill, alloc-stall, scan-async, scan-direct, reclaim-async, and reclaim-direct.
[PATCH] Documentation: coding-style: don’t encourage WARN*() (Alex Elder) https://lkml.kernel.org/r/20240414170850.148122-1-elder@linaro.org
Due to the panic_on_warn setting, even WARN() and its friends are not that welcomed in multiple situations. Alex posted a patchset to make this point clearer.
[PATCH v4 0/9] transfer page to folio in KSM (Alex Shi) https://lkml.kernel.org/r/20240409092826.1733637-1-alexs@kernel.org
Alex sent the first patchset for making KSM uses folios.
[PATCH] tracing: Add new_exec tracepoint (Marco Elver) https://lkml.kernel.org/r/20240408090205.3714934-1-elver@google.com
Marco sent a patch for adding a new tracepoint. The tracepoint is for tracing a case where new exec starts while having original state including the mm.
Announcing Community Engagement Working Group (Nikolai Kondrashov) https://lkml.kernel.org/r/bf81be70-61ec-4169-b66a-5c3136869107@gmail.com/
KernelCI project announces Community Engagement Working Group which will purpose to connect with kernel maintainers and discuss test quality improvement for their subsystems.
[PATCH v3 0/7] mm/kvm: Improve parallelism for access bit harvesting (James Houghton) https://lkml.kernel.org/r/20240401232946.1837665-1-jthoughton@google.com
James posted the third version of the patchset for adding KVM’s test-and-clear-access-bits fast path that doesn’t take mmu_lock. It is RFC because the author wants to do more performance tests, and unusre if the arm64 part is correct.
[PATCH] mm, mmap: remove vma_merge() (Vlastimil Babka) https://lkml.kernel.org/r/20240401192623.18575-2-vbabka@suse.cz
After the successful removal of SLAB, Vlastimil suggests to remove vma_merge(). But the patch has sent on 2024-04-01 :)
[PATCH 0/2] support multi-size THP numa balancing (Baolin Wang) https://lkml.kernel.org/r/cover.1711453317.git.baolin.wang@linux.alibaba.com
Baolin posted a patchset for making NUMA balancing algorithm for multi-sized THP (mTHP) to follow the THP strategy as the basic support.
[PATCH v2 0/6] Improve visibility of writeback (Kemeng Shi) https://lkml.kernel.org/r/20240327155751.3536-1-shikemeng@huaweicloud.com
Kemeng posted a patchset that adds two new debugfs file, bdi/xxx/stats and bdi/xxx/wb_stats to show whole bdi and per-wb writeback information.
[PATCH v10] zswap: replace RB tree with xarray (Chris Li) https://lkml.
[GIT PULL] slab updates for 6.9 (Vlastimil Babka) https://lkml.kernel.org/r/c021631d-29e5-432f-bfcd-1d75ec28d065@suse.cz
Vlastimil sent the slab pull request for 6.9 to Linus Torvalds.
[PATCH v4 0/6] Swap-out mTHP without splitting (Ryan Roberts) https://lkml.kernel.org/r/20240311150058.1122862-1-ryan.roberts@arm.com
Ryan sent the fourth version of the patchset to swap-out multi-sized THP without splitting first. The approach is similar to that for PMD-sized THP.
[PATCH] vmstat: Keep count of the maximum page reached by the kernel stack (Pasha Tatashin) https://lkml.kernel.org/r/20240313033417.447216-1-pasha.tatashin@soleen.com
[RFC 01/16] mm/hmm: let users to tag specific PFNs (Leon Romanovksy) https://lkml.kernel.org/r/a77609c9c9a09214e38b04133e44eee67fe50ab0.1709631413.git.leon@kernel.org
This RFC patchset introduces new flag that allows users to tag specific PFNs with extra data.
[PATCH V3 01/10] mm: page_alloc: freelist migratetype hygiene (Johannes Weiner) https://lkml.kernel.org/r/20240306041526.892167-1-hannes@cmpxchg.org
Whne pages of one migratetype end up on the freelists on another type, migratetype violations happen. This third version of patchset fixes it.
[PATCH 0/8] tracing: Persistent traces across a reboot or crash (Steven Rostedt) https://lkml.
[PATCH 1/1] mm/madvise: enhance lazyfreeing with mTHP in madvise_free (Lance Yang) https://lkml.kernel.org/r/20240225123215.86503-1-ioworker0@gmail.com
Lance improves madvise_free_pte_range() to handle large folios that smaller than PMD-size and optimizes lazy-free of PTE-mapped smaller than PMD-size THPs using this patch.
[RFC PATCH] mm: Add reclaim type to memory.reclaim (Yafang Shao) https://lkml.kernel.org/r/20240225114204.50459-1-laoar.shao@gmail.com
This RFC patch extends memory.reclaim memcg file to receive type of memory to reclaim. It supports pagecache only, slabs only, and both.
[LSF/MM/BPF TOPIC] mm/mempolicy extentions for heterogeneous memory systems (Gregory Price) https://lkml.
[PATCH V2 0/2] Introduce slabinfo version 2.2 (Fangzheng Zhang) https://lkml.kernel.org/r/20240219031911.10372-1-fangzheng.zhang@unisoc.com
This patchset updates slabinfo to version 2.2 and adds slabreclaim column for recording whether each slab pool is reclaim type.
[RFC PATCH 0/3] make the hugetlb migration strategy consistent (Baolin Wang) https://lkml.kernel.org/r/cover.1708507022.git.baolin.wang@linux.alibaba.com
Falling back to other NUMA nodes when migrating freed hugetlb is prevented while it is allowed for in-use hugetlb. This patchset make it more clear and consistent.
[Invitation] Linux MM Alignment Session on Cold Page Detection on Wednesday (David Rientjes) https://lkml.
[PATCH v3] cgroup2: New memory.max.effective like cgroup1 hierarchical_memory_limit (Jan Kratochvil) https://lkml.kernel.org/r/ZctiM5RintD_D0Lt@host1.jankratochvil.net
Unlike cgroup v1, which was able to find effective memory limit using hierarchical_memorylimit and hierarchical_memsw_limit lines, cgoup v2 doesn’t have such hierarchical information, so need to read all files in the hierarchy to know the effective memory limit. This patchset adds memory.max.effective and memory.swap.max.effective to let users easily get the information.
[PATCH v3 00/35] Memory allocation profiling (Suren Baghdasaryan) https://lkml.