Posts

[RFC 0/2] mm: introduce THP deferred setting (Nico Pache) https://lore.kernel.org/20240729222727.64319-1-npache@redhat.com Nico found increased memory footprint from THP after their customers switching from RHEL7 to RHEL8. To mitigate the issue, Nico is proposing adding a new THP configuration called deferred. It is same to madvise but allows khugepaged to do the work. [PATCH v2 0/4] mm: clarify nofail memory allocation (Barry Song) https://lore.kernel.org/20240731000155.109583-1-21cnbao@gmail.com Barry posted the second version of their patchset for __GFP_NOFAIL clarification....

[PATCH 00/17] mm: introduce numa_memblks (Mike Rapoport) https://lore.kernel.org/20240716111346.3676969-1-rppt@kernel.org Mike Rapoport posted thi spatchset for moving numa_memblks from x86 to the generic code, as a followup of the discussion for CXL memory windows handling on arm64. [RFC PATCHv2] mm: introduce reclaim throttle in MGLRU (Zhaoyang Huang) https://lore.kernel.org/20240716094348.2451312-1-zhaoyang.huang@unisoc.com Zhaoyang sent a patch that makes direct reclaim of MGLRU be throttled by judging the number of isolated and inactive folios, to prevent issues similar to classic LRU....

I just made a DAMON logo using DAMON, like below. $ git clone https://github.com/sjp38/masim && cd masim $ cat damon_pixel_2 11111111 11 11 111111 11111111 11 11 11111111 11111111 1111 11111111 11111111 11 11 11111111 11111111 1111 11111111 $ ./pixels_to_access_config.py ./damon_pixel_2 $((100*1024*1024)) 300 damon.cfg $ sudo damo record "./masim ./configs/stairs.cfg" $ sudo damo report heatmap --output damon.png The output is below: The cropped one:

[RFC][PATCH 0/4] slab: Allow for type introspection during allocation (Kees Cook) https://lore.kernel.org/20240708190924.work.846-kees@kernel.org Kees posted this RFC patch, which allows kmalloc() be able to know the type of the object that it is requested to allocate memory for. It is done by letting kmalloc() optionally receives the pointer instead of the size. [Invitation] Linux MM Alignment Session on guest_memfd() mapping, hugepages, and page pinning on Wednesday https://lore.kernel.org/197a2f19-c71c-fbde-a62a-213dede1f4fd@google.com David invites MM community to another instance of the MM alignment....

[RFC PATCH 0/4] Introduce PMC(PER-MEMCG-CACHE) (Huan Yang) https://lore.kernel.org/20240702084423.1717904-1-link@vivo.com Huan posted an RFC for per-memcg-cache, which is for utilizing per-memcg cache of certain rules. The goal is allowing memory that released by an application be allocated with high priority within the application, to improve locality and performance while avoiding unnecessary memory reclamation. [PATCH] mm: remove CONFIG_MEMCG_KMEM (Johannes Weiner) https://lore.kernel.org/20240701153148.452230-1-hannes@cmpxchg.org Since CONFIG_MEMCG_KMEM was default-enabled and equivalent to CONFIG_MEMCG for almost a decade, Johannes tries to remove the config option....