Linux 7.4 will be the first cycle to draw core memory management changes from a newly agreed "Core MM" merge tree, splitting them away from the long-standing mixed-traffic MM branch maintained by Andrew Morton, according to Phoronix.
The reorganisation is purely a change to how patches travel through the kernel's integration hierarchy — it affects routing, not how memory management behaves once the code is in-tree.
What is changing
For years, core MM work has reached mainline through Andrew Morton's "MM" branch, a route that also carried a much wider mix of memory-management-adjacent material — most notably the steady stream of DAMON and DAMOS (data access monitoring and optimization) patches, along with assorted subsystem submissions that happened to queue on the same branch.
Under the new arrangement, agreed among kernel maintainers, all core memory management changes get their own dedicated "Core MM" tree as their home going forward. DAMON/DAMOS and the rest of the broader traffic stay on the existing route, so the two streams now merge into mainline separately rather than as one bundle.
The practical result is cleaner patch provenance. Reviewers looking specifically at core MM no longer have to pick those changes out of a mixed queue, and the boundary between "core MM" and "everything else memory-related" becomes explicit in the merge graph.
Why it matters downstream
The likely practical effects for downstream maintainers — distribution integrators and stable-tree maintainers in particular — are easier identification of Core MM fixes for targeted backports, since they no longer arrive entangled with unrelated work on a shared branch; a narrower stream to bisect when a memory-management regression is suspected mid-cycle; and clearer visibility during the merge window over which portion of MM traffic has landed.
None of this affects machines already in production; the benefit shows up when maintainers assemble and debug future kernels. The story is essentially one of scale: the MM subsystem has grown large enough that a single integration branch carrying mixed traffic started to impede review, and the split offers a template that could be applied to other overburdened branches if it proves out.
Cycle context
The MM reshuffle is a plumbing change rather than a feature, but it is the kind most likely to be noticed by anyone tracking merge paths as 7.4 moves toward release.
What to watch
The open question is whether the boundary between Core MM and the existing MM branch stays stable, or creeps as edge-case patches get routed inconsistently.
For now, the change is a workflow story, not a technical one — but it is the kind of workflow change that quietly shapes how quickly memory-management fixes reach the kernels that administrators actually run.
Source: Phoronix, "Linux 7.4 Begins Employing New 'Core MM' Tree".
據 Phoronix 報道，Linux 7.4 將成為首個從新商定的「Core MM」merge tree 取得核心記憶體管理改動的版本，並將這部分 patch 與一直以來由 Andrew Morton 維護、混雜多類內容的 MM 分支分開處理。
這次重整純粹是 patch 穿越 kernel 整合層級方式的變動——影響的是分流方式，而非程式碼進入 tree 後記憶體管理的運作行為。
有何改變
多年以來，核心 MM 工作一直是透過 Andrew Morton 的「MM」分支進入 mainline。該分支同時承載種類龐雜、與記憶體管理相鄰的材料——其中最顯著的是源源不絕的 DAMON 及 DAMOS（data access monitoring and optimization）patch，以及剛好排在同一分支的各類子系統提交。
根據 kernel 維護者達成的新安排，所有核心記憶體管理改動今後將以專用的「Core MM」tree 作為歸屬。DAMON／DAMOS 及其餘較廣泛的內容則維持原有路線，兩條流如今各自獨立 merge 進入 mainline，而不再捆綁成一體。
實際結果是 patch 的出處更為清晰。專門審閱核心 MM 的維護者毋須再從混雜的 queue 中逐一挑出相關改動，而「核心 MM」與「其他一切記憶體相關事項」的界線在 merge graph 中亦變得明確。
對下游的影響
對下游維護者——尤其是各發行版整合者及 stable tree 維護者——可能產生的實際影響包括：由於 Core MM 修正不再與共享分支上毫不相干的工作糾纏在一起，更易識別並進行針對性的 backport；當懷疑周期中途出現記憶體管理回歸時，可供二分偵錯（bisect）的範圍更窄；以及在 merge window 期間能更清楚掌握哪部分 MM 內容已經落地。
以上一切均不影響已在生產環境運作的機器；相關好處將於維護者日後組裝及偵錯 kernel 時浮現。這個故事的本質在於規模：MM 子系統已增長至單一整合分支承載混合流量開始妨礙審閱的程度，而這次拆分提供了一個範本，一旦證明可行，日後或可套用於其他負荷過重的分支。
版本周期背景
MM 這次重組屬於 plumbing 層面的改動，並非功能更新，但隨著 7.4 逐步走向發布，它正是追蹤 merge 路徑人士最可能注意到的那類變動。
值得留意之處
目前未有定論的是，Core MM 與現有 MM 分支之間的界線能否維持穩定，抑或會因邊界情況的 patch 分流不一而逐漸模糊。
暫時而言，這是一則工作流程的新聞，而非技術新聞——但它正是那類會悄然左右記憶體管理修正到達管理員實際運行 kernel 之速度的工作流程變動。
來源：Phoronix，〈Linux 7.4 Begins Employing New 'Core MM' Tree〉。