[PATCH] x86/mm: drop 4MB restriction on minimal NUMA node size (Mike Rapoport) https://lkml.kernel.org/r/20231017062215.171670-1-rppt@kernel.org
Kernel can panic when less than 4MB NUMA node is configured. The commit which introduced the 4MB limit doesn’t clearly explain the reason. This patch drops the limit.
[Invitation] Linux MM Alignment Session on Small-sized THPs on Wednesday (David Rientjes) https://lkml.kernel.org/r/76c2f454-5ab2-61db-a41f-a30f08a4aa1e@google.com
The next MM alignment session sill be lead by Ryan Roberts from ARM, for small-sized THP.
[RFC PATCH 0/5] hot page swap to zram, cold page swap to swapfile directly (Lincheng Yang) https://lkml.kernel.org/r/20231008095924.1165106-1-lincheng.yang@transsion.com
The team has made the kernel feature that swaps hot page to zram and cold pages to swapfile, directly. It aims to reduce zram compression/decompression overhead under memory pressure on Android v4.19 linux kernel, and trying to make it upstreamed due to the GKI limitation.
[RFC PATCH] zswap: add writeback_time_threshold interface to shrink zswap pool (Zhongkun He) https://lkml.
[PATCH 0/5] riscv: Use READ_ONCE()/WRITE_ONCE() for pte accesses (Alexandre Ghiti) https://lkml.kernel.org/r/20231002151031.110551-1-alexghiti@rivosinc.com
A patchset for making all page table entries accesses use READ_ONCE()/WRITE_ONCE(), to avoid compiler reordering of the accesses.
[PATCH v1 0/3] mm/gup: Introduce pin_user_pages_fd() for pinning shmem/hugetlbfs file pages (Vivek Kasireddy) https://lkml.kernel.org/r/20231003074447.3245729-1-vivek.kasireddy@intel.com
A patchset introducing a new function for long-term pages pinning.
[RFC PATCH v3 0/3] Introduce persistent memory pool (Stanislav Kinsburskii) https://lkml.kernel.org/r/169645773092.11424.7258549771090599226.stgit@skinsburskii.
An RFC implementing a memory allocator in the kernel for persistent memory has posted.
[PATCH v3 0/4] Smart scanning mode for KSM (Stefan Roesch) https://lkml.kernel.org/r/20230926040939.516161-1-shr@devkernel.io
KSM scans pages without historic information, so pages that already found to be not eligible for deduplication are scanned unnecessarily. This patchset introduces smart scanning for KSM, which utilizes the historic information.
[PATCH -V2 00/10] mm: PCP high auto-tuning (Huang Ying) https://lkml.kernel.org/r/20230926060911.266511-1-ying.huang@intel.com
This patchset implements auto-tuning of Per-CPU Pageset high to optimize page allocation performance.
[PATCH rfc 0/5] mm: improve performance of kernel memory accounting (Roman Gushchin) https://lkml.
[Invitation] Linux MM Alignment Session on APIs for Sparse Mappings on Wednesday (David Rientjes) https://lkml.kernel.org/r/fba78307-e4a3-2621-2917-464f19bb4c57@google.com
David is organizing yet another Linux MM alignment session for proper API for sparse mappings.
[PATCH v2] vmscan: add trace events for lru_gen (Jaewon Kim) https://lkml.kernel.org/r/20230921062206.14429-1-jaewon31.kim@samsung.com
This patch adds two new tracevents called trace_mm_vmscan_lru_gen_scan and trace_mm_vmscan_lru_gen_evict that similar to trace_mm_vmscan_lru_isolate and trace_mm_vmscan_lru_shrink_[in]active.
[PATCH v3 00/15] futex: More futex2 bits (Peter Zijlstra) https://lkml.kernel.org/r/20230921104505.717750284@noisy.programming.kicks-ass.net
Peter sent third version of the patchset that introduces a new interface for same futex core, futex2.
[syzbot] Monthly mm report (Sep 2023) https://lkml.kernel.org/r/000000000000c06b600605127cb3@google.com
Syzbot posted its mm subsystem report for last 31 days. For the period, seven new issues were found and two among those were fixed. In total, 43/232 issues are still open.
[PATCH 0/2] workload-specific and memory pressure-driven zswap writeback (Nhat Pham) https://lkml.kernel.org/r/20230911164024.2541401-1-nphamcs@gmail.com
Zswap has only single LRU list and shrinks under user-defined limit. This patch series seperates the LRU list out to per-memcg and per-NUMA, and perform workload-specific writeback under memory pressure.
[GIT PULL] MM updates for 6.6-rc1 (Andrew Morton) https://lkml.kernel.org/r/20230828184251.bfbc164280503903eacf498c@linux-foundation.org
Andrew sent the mm subsystem pull request for v6.6-rc1 to Linus Torvalds. And, it pulled.
[PATCH 1/2] mm/vmalloc: Add a safer version of find_vm_area() for debug (Joel Fernendes) https://lkml.kernel.org/r/20230830110402.386898-1-joel@joelfernandes.org
A patchset implementing safe trylock version of vmalloc dump function.
[PATCH v2 0/6] Introduce __mt_dup() to improve the performance of fork() (Peng Zhang) https://lkml.kernel.org/r/20230830125654.21257-1-zhangpeng.00@bytedance.com
When doing fork(), vma of the parent should be inserted into the maple tree of the child.
[PATCH] maple_tree: shrink struct maple_tree from 24 to 16 bytes on LP64 (Mateusz Guzik) https://lkml.kernel.org/r/20230821225145.2169848-1-mjguzik@gmail.com
This patchset moves ma_root field to be placed after ma_flags to plug a padding hole of struct mape_tree, to reduce the size of the struct.
idea for scalable and almost waste-free percpu counter alloc (was: Re: [PATCH 0/2] execve scalability issues, part 1) (Mateusz Guzik) https://lkml.kernel.org/r/CAGudoHH-b5R+COHUtns92NTzSVeT1ECMHb22SEcH_108Pi0FTQ@mail.gmail.com
Mateusz is suggesting another idea for scalable and almost-waste-free percpu counter alloc.
[MAINTAINERS SUMMIT] Handling of embargoed security issues – security@korg vs. linux-distros@ (Jiri Kosina) https://lkml.kernel.org/r/nycvar.YFH.7.76.2308150927190.14207@cbobk.fhfr.pm
Yet another maintainers summit agenda proposal about security issue handling has posted. There are securitiy@korg and linux-distros@. The kernel community wants to do handling of the security fixes from only security@, without being bothered by linux-distros@. Jiri proposes to discuss about if it could eventually better handled with linux-distros@.
Maintainers Summit 2023 Call for Topics (Theodore Ts’o) https://lkml.
[RFC PATCH 0/3] Reduce NUMA balance caused TLB-shootdowns in a VM (Yan Zhao) https://lkml.kernel.org/r/20230808071329.19995-1-yan.y.zhao@intel.com
For virtual machines having assigned devices or VFIO mediated devices, all or part of its memory are pinned for long-term. Auto NUMA balancing can cause unnecessary TLB-shootdowns. This RFC patchset proposes a solution for that.
[PATCH v2] cma: introduce CMA_ALLOC_DEBUG config (Bibek Kumar Patro) https://lkml.kernel.org/r/20230809131640.18791-1-quic_bibekkum@quicinc.com
CONFIG_CMA_DEBUG makes all CMA debugging message enabled. This can results in flooding the dmesg buffer.