linux

My opensource commits statistics in 2023

As today is the last day of 2023, I checked some open source commits statistics using my simple and buggy script, and github. Similar statistics for 2022 is available at another post. In short, I made 8th and 4th biggest changes to Linux kernel memory management subsystem among the 295 people, for lines of changes (1,910) and commits (66). For the Linux kernel whole tree, the numbers become 264th (3,562 lines) and 80th (147 commits) among 5,006 people.

LKML News v6.7-rc8

[PATCH] kernel: Introduce a write lock/unlock wrapper for tasklist_lock (Maria Yu) https://lkml.kernel.org/r/20231225081932.17752-1-quic_aiquny@quicinc.com A patch introducing a write lock/unlock wrapper for tasklist_lock, to overcome bad responsive performance of system that coming from scenarios that acquiring read lock which wrote lock needed to be waiting for. [PATCH v3 00/10] IOMMU memory observability (Pasha Tatashin) https://lkml.kernel.org/r/20231226200205.562565-1-pasha.tatashin@soleen.com Third version of patchset to allow monitoring and limiting IOMMU memory usage is posted. [GIT PULL] hotfixes for 6.

LKML News v6.7-rc7

[PATCH v5 0/4] mm/ksm: Add ksm advisor (Stefan Roesch) https://lkml.kernel.org/r/20231218231054.1625219-1-shr@devkernel.io Stefan posted the fifth version of the patchset for automatic and efficient tuning of kernel samepage merging. [PATCH -next v4 0/2] Make memory reclamation measurable (Bixuan Cui) https://lkml.kernel.org/r/20231220012123.2787-1-cuibixuan@vivo.com This patchset adds a new tracer to shrink_active_list(), shrink_inactive_list(), and shrinker->count_objects() to make memory recycling mesaurable with bpf. [RFC PATCH 1/1] mm: mark folio accessed in minor fault (Zhaoyang Huang) https://lkml.kernel.org/r/20231220102948.1963798-1-zhaoyang.huang@unisoc.com Inactive mapped folio is promoted to active only when it is scanned in shrink_inactive_list().

LKML News v6.7-rc6

[Invitation] Linux MM Alignment Session on remap_file_pages on Wednesday (David Rientjes) https://lkml.kernel.org/r/46f97bec-3a48-e714-0a8d-b5206e1704ba@google.com There will be the last Linux MM Alignment Session for re-introducing remap_file_pages(), led by Matthew Wilcox, on 2023-11-13. [PATCH] kernel: Introduce a write lock/unlock wrapper for tasklist_lock (Maria Yu) https://lkml.kernel.org/r/20231213101745.4526-1-quic_aiquny@quicinc.com There are multiple scenarios for tasklist_lock that could lead to bad responsiveness. To mitigate the situation, this patch introduces a write lock/unlock wrapper for tasklist_lock. [PATCH 0/9] Finish two folio conversions (Matthew Wilcox) https://lkml.

LKML News v6.7-rc5

[PATCH] mm/thp: add CONFIG_TRANSPARENT_HUGEPAGE_NEVER option (Dmytro Maluka) https://lkml.kernel.org/r/20231204163254.2636289-1-dmaluka@chromium.org Khugepaged has its own memory cost, and it increases vm.min_free_kbytes as it needs when it starts. On 4GB RAM, 3 mmzones and pageblock_order == MAX_ORDER, khugepaged increases vm.min_free_kbytes from 8MB to 132MB. This patch introduces a config to disallow khugepaged starts by default. It still allows the userspace to start khugepaged, though. [PATCH v4 0/2] KSM: support smart-scan feature (Stefan Roesch) https://lkml.kernel.org/r/20231205180952.3843711-1-shr@devkernel.io

LKML News v6.7-rc4

[PATCH v2] prctl: Get private anonymous memory region name (Rong Tao) https://lkml.kernel.org/r/tencent_77E0BC1E8E9A21CA1B9009DD66555C72A90A@qq.com Now prctl() can set the name of anonymous vmas, but there is no option for getting the name. This patch implements the getter feature. [RFC PATCH 0/6] Supporting GMEM (generalized memory management) for external memory devices (Weixi Zhu) https://lkml.kernel.org/r/20231128125025.4449-1-weixi.zhu@huawei.com Because Linux kernels’ memory management system is mainly for only host DRAM, accelerator driver developers need to invent their own external memory management subsystems that feature similar functionality.

LKML News v6.7-rc3

[PATCH v6 0/4] page_owner: print stacks and their counter (Oscar Salvador) https://lkml.kernel.org/r/20231120084300.4368-1-osalvador@suse.de This patchset introduces a new read-only file page_owner_stacks. This can be useful for getting the relationship between stacks and allocation/free without fiddling with pages. [PATCH v3 0/8] riscv: ASID-related and UP-related TLB flush enhancements (Samuel Holland) https://lkml.kerenl.org/r/20231122010815.3545294-1-samuel.holland@sifive.com A patchset for improving some TLB flush inefficiencies in risc-v. [RFC] mm: Multi-gen LRU: fix use of mm/page_idle/bitmap (Henry Huang) https://lkml.kernel.org/r/cover.1700719508.git.henry.hj@antgroup.com

LKML News v6.7-rc2

[PATCH v7 00/10] Small-sized THP for anonymous memory (Ryan Roberts) https://lkml.kernel.org/r/20231115132734.931023-1-ryan.roberts@arm.com The seventh revision of small-sized anonymout memory THP is posted. [PATCH v2 00/14] Transparent Contiguous PTEs for User Mappings (Ryan Roberts) https://lkml.kernel.org/r/20231115163018.1303287-1-ryan.roberts@arm.com This patchset makes opportunistic and transparent contpte mappings for user memory when it meets the requirement. [GIT PULL] hotfixes for 6.7-rc2 (Andrew Morton) https://lkml.kernel.org/r/20231117140848.02adcf56c301bb3ab85ae75a@linux-foundation.org MM subsystem hotfixes for rc2 has sent to Linus Torvalds. Linux 6.7-rc2 (Linus Torvalds) https://lkml.

LKML News v6.7-rc1

NUMA + THP test suites? (John Hubbard) https://lkml.kernel.org/r/21812f89-c1a0-4ca0-9f73-54f20c4d29ef@nvidia.com John asks tets suite for NUMA and THP that can be useful for testing patches for the topics. [PATCH v4 0/1] mm: report per-page metadata information (Sourav Panda) https://lkml.kernel.org/r/20231031223846.827173-1-souravpanda@google.com This patch adds per-node and global page metadata information to sysfs and procfs, to let users know how much memory is being used by the metadata. [GIT PULL] MM updates for 6.7-rc1 (Andrew Morton) https://lkml.

LKML News v6.6

Virtual Scanning Considered Harmful (Matthew Wilcox) https://lkml.kernel.org/r/ZTc7SHQ4RbPkD3eZ@casper.infradead.org Matthew shares his benchmark program that simulates speeds of LRU pages traversal using an linked list, an array, and direct physical addresses. As can expect, the physical address based traversal is fastest. [GIT PULL] hotfixes for 6.6 (Andrew Morton) https://lkml.kernel.org/r/20231024094608.f2dc0ca5ed2619d411a31df5@linux-foundation.org Andrew sent mm hotfixes for v6.6 to Linus Torvalds. One for DAMON also contained. [PATCH v4 0/5] workload-specific and memory pressure-driven zswap writeback (Nhat Pham) https://lkml.