LKML News v6.5-rc6

[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. This patch introduces another debugging config that enable only cma bit allocation status.

[PATCH] mm: sparse: shift operation instead of division operation for root index (Gui Hui)

https://lkml.kernel.org/r/20230810103829.10007-1-guohui@uniontech.com

This patch makes __nr_to_section() to use shift operation instead of division operation to improve the performance.

[syzbot] Monthly mm report (Aug 2023) (syzbot)

https://lkml.kernel.org/r/0000000000006518110602a0e0c0@google.com

Syzbot summarizes new and fixed bugs that it find from mm subsystem for last 31 days. One new issue has found but no bug was fixed. In total, 40 issues are still open.

[MAINTAINER SUMMIT] coping with stress as a maintainer (James Bottomley)

https://lkml.kernel.org/r/ab9cfd857e32635f626a906410ad95877a22f0db.camel@HansenPartnership.com

James proposes a topic for the kernel maintainer summit of this year to discuss about stress of maintainer works.

[GIT PULL] hotfixes for 6.5-rc6 (Andrew Morton)

https://lkml.kernel.org/r/20230811134809.5f1e6cfc690baaf796ff4072@linux-foundation.org

Andrew Morton sent a pull request for mm hotfixes to Linus Torvalds.

Linux 6.5-rc6 (Linus Torvalds)

https://lkml.kernel.org/r/CAHk-=wg4pV_H8bWsfWPQGkcPUyHdVApc6pf6eETkZAF_zxgNNw@mail.gmail.com

The AMD h/w mitigation patch was somewhat might looks like outstanding for the last week. Torvalds says other than that, everything looks normal, and even argues h/w mitigations should be treated as normal, though such fixes would require additional fixes for the fixes, mainly because those initial fixes are usually made under embargo, without many eyeballs.

Below is the diffstat of the releases in the last two years.

Note that the y-axis is in logarithm. I draw it using https://github.com/sjp38/relstat and https://github.com/sjp38/lazybox using below command:

$ relstat.py --since 2021-08-14 | ~/lazybox/gnuplot/plot.py \
    --data_fmt table --type labeled-lines --xtics_rotate -90 \
    --font "Times New Roman, 5pt" --ylog --pointsize 0.3

And, below is the diffstat of the -rc6 releases in the last two years.