lkml news

LKML News v6.0-rc6

[RFC PATCH] mm: track bad page via kmemleak (Zhaoyang Huang) https://lkml.kernel.org/r/1663679468-16757-1-git-send-email-zhaoyang.huang@unisoc.com This RFC patch makes bad pages to be tracked via kmemleak. [PATCH v1 0/3] coding-style.rst: document BUG() and WARN() rules (David Hildenbrand) https://lkml.kernel.org/r/20220920122302.99195-1-david@redhat.com There was a recent discussion on when to use BUG() and WARN() and the variants. This patch tries to make it clearly described in coding-style doc and checkpatch script. [RFC 0/6] migrate_pages(): batch TLB flushing (Huang Ying) https://lkml.

LKML News v6.0-rc6

[PATCH mm-unstable v15 00/14] Multi-Gen LRU Framework (Yu Zhao) https://lkml.kernel.org/r/20220918080010.2920238-1-yuzhao@google.com A new version of MGLRU patchset has posted. Among the changes in this new version, fix of long-tailed direct reclaim latency looks interesting. Linux 6.0-rc6 (Linus Torvalds) https://lkml.kernel.org/r/CAHk-=wi=gtuSO8Yz8LDubkMk7TiMsZxpypt9S10jeZRkyaBFnA@mail.gmail.com The sixth release candidate for Linux v6.0 has out. Maybe due to the LPC/OSS EU/Maintainers summit which all were held in last week, this release is quite small. Torvalds says it might mean that seventh release will be bigger than usual and therefore it might need one more week for v6.

LKML News v6.0-rc5

[LPC 2022] Kernel Memory Management Microconference - next Monday! (Vlastimil Babka) https://lkml.kernel.org/r/e918dd3c-aa71-c510-4a50-4f9882bca3e8@suse.cz Vlastimil and Matthew are reminding about the kernel memory management microconference, which will be held as a part of the LPC this year, in Dublin. [PATCH v6 00/44] Add KernelMemorySanitizer infrastructure (Alexander Potapenko) https://lkml.kernel.org/r/20220905122452.2258262-1-glider@google.com Sixth iteration of the KMSAN infra patchset has posted. [PATCH v3 0/2] Introduce CABA helper process tree (Pavel Tikhomirov) https://lkml.kernel.org/r/20220908140313.313020-1-ptikhomirov@virtuozzo.com Current linux process tree looses the information in which subtree the child was originally born, after the children is died and therefore reparented.

LKML News v6.0-rc4

[RFC PATCH 00/28] per-VMA locks proposal (Suren Baghdasaryan) https://lkml.kernel.org/r/20220829212531.3184856-1-surenb@google.com The VMA lock scalability issue is still there. This RFC patchset shows the proof of the concept for the per-VMA lock idea, which was discussed in LSF/MM this year, as part of the SPF discussion. [RFC PATCH 00/30] Code tagging framework and applications (Suren Baghdasaryan) https://lkml.kernel.org/r/20220830214919.53220-1-surenb@google.com This RFC patchset introduces code tagging framework which can identify a specific location in the source code when compile.

LKML News v6.0-rc3

[mm] 96db82a66d: will-it-scale.per_process_ops -95.3% regression (kernel test robot) https://lkml.kernel.org/r/YwIoiIYo4qsYBcgd@xsang-OptiPlex-9020 A big regression of mm multi-cpu scalability test (will-it-scale) has found by the kernel test robot. The commit 96db82a66d1db0e2966c02dc492e02eeb3a94a75 (“mm: align larger anonymous mappings on THP bundaries”) is pointed out as the suspect. [PATCH v13 00/70] Introducing the Maple Tree (Liam Howlett) https://lkml.kernel.org/r/20220822150128.1562046-1-Liam.Howlett@oracle.com The thirteenth version of the maple tree patchset has posted after rebased on latest mm-unstable and fixed some bugs.

LKML News v6.0-rc2

[PATCH v14 00/14] Multi-Gen LRU Framework (Yu Zhao) https://lkml.kernel.org/r/20220815071332.627393-1-yuzhao@google.com Fourteenth version of MGLRU patch has rebased on the latest mm-unstable and posted again. This may be merged in the mm tree. folio_map (Matthew Wilcox) https://lkml.kernel.org/r/YvvdFrtiW33UOkGr@casper.infradead.org Matthew explains his proposal on folio_map in detail and asks if others have better ideas. [PATCH v3] mm: add thp_utilization metrics to debugfs (Alexander Zhu) https://lkml.kernel.org/r/20220818000112.2722201-1-alexlzhu@fb.com This patch adds debugfs file for tracking how THP is utilized well.

LKML News v6.0-rc1

[PATCH] mm: vmscan: fix extreme overreclaim and swap floods (Johannes Weiner) https://lkml.kernel.org/r/20220802162811.39216-1-hannes@cmpxchg.org Johannes found proactive reclamation using memory.reclaim sometimes reclaim more pages than requested on specific cgroup use cases. This patch resolves the issue so that the proactive reclamation doesn’t over-reclaim pages. [RFC PATCH 0/4] propose a auto-run mode for ksm (Xu Xin) https://lkml.kernel.org/r/20220803100306.1653382-1-xu.xin16@zte.con.cn KSM operation management is quite manual and sometimes inefficient if not well managed. This patchset introduces auto-run mode that can operate KSM in a better way.

LKML News v5.19

[PATCH v4 0/3] mm/mprotect: Fix soft-dirty checks (Peter Xu) https://lkml.kernel.org/r/20220725142048.30450-1-peterx@redhat.com Fourth version of the soft-dirty check fixes for mprotect has posted. This version polishes the commit message and adds tests. [PATCH v2 1/2] mm/slub: enable debugging memory wasting of kmalloc (Feng Tang) https://lkml.kernel.org/r/20220725112025.22625-1-feng.tang@intel.com As slab allocates memory in only power of 2 bytes size, memory can be wasted when slab allocation is requested for a large number of times for objects that sizes not aligned in power of two bytes.

LKML News v5.19-rc8

[RESEND PATCH v3] arm64: enable THP_SWAP for arm64 (Barry Song) https://lkml.kernel.org/r/20220718090050.2261-1-21cnbao@gmail.com This patch enables THP_SWAP which can improve the swap performance for transparent hugepages on ARM64 architecture. [PATCH v12 00/69] Introducing the Maple Tree (Liam Howlett) https://lkml.kernel.org/r/20220720021727.17018-1-Liam.Howlett@oracle.com Maple tree has already merged in the mm-unstable but a new bug has found, so the 12th version of the patchset has posted. [PATCH v10 0/8] mm/demotion: Memory tiers and demotion (Aneesh Kumar K.

LKML News v5.19-rc7

[PATCH 0/4] add struct page and Direct I/O support to reserved memory (Li Chen) https://lkml.kernel.org/r/20220711122459.13773-1-me@linux.beauty Only buffered I/O is available to reserved memory for ZONE_DEVICE and mhp. This patchset enables direct I/O to the memory for better throughput. [PATCH -V4 RESEND 0/3] memory tiering: hot page selection (Huang Ying) https://lkml.kernel.org/r/20220713083954.34196-1-ying.huang@intel.com This patchset implements a new hot page identification algorithm based on the latency between NUMA balancing page table scanning and hit page fault, so that it can be somewhat LFU algorithm instead of MRU algorithm.