LKML News v5.10-rc1

[PATCH] mm: Make allocator take care of memoryless numa node (Xianting Tian)

https://lkml.kernel.org/r/20201012082739.15661-1-tian.xianting@h3c.com

CPUs in powerpc like architecture could have no local memory. As the kernel code doesn’t care about it, this commit makes it to handle the case.

incoming (Andrew Morton)

https://lkml.kernel.org/r/20201013164658.3bfd96cc224d8923e66a9f4e@linux-foundation.org

This is the pull request from Andrew Morton, the mm maintainer, for v5.10.

[PATCH v4 0/5] Speed up mremap on large regions (Kalesh Singh)

https://lkml.kernel.org/r/20201014005320.2233162-1-kaleshsingh@google.com)

Repost fo the patch series. Change from v3 is only adding Kirill’s Acked-by’s.

incoming (Andrew Morton)

https://lkml.kernel.org/r/20201015194043.84cda0c1d6ca2a6847f2384a@linux-foundation.org

Another pull request from Andrew Morton. Changes for THP, which removes the assumption on THP size seems interesting.

incoming (Andrew Morton)

https://lkml.kernel.org/r/20201017161314.88890b87fae7446ccc13c902@linux-foundation.org

Yet another pull request from Andres Morton. Minchan Kim’s process_madvise() patchset is in here.

Splitting a THP beyond EOF (Matthew Wilcox)

https://lkml.kernel.org/r/20201020014357.GW20115@casper.infradead.org

Matthew shares a weird corner case that THP is used beyond the end of file and his idea on the solution of this problem.

[PATCH] mm: don’t wake kswapd prematurely when watermark boosting is disabled (Johannes Weiner)

https://lkml.kernel.org/r/20201020175833.397286-1-hannes@cmpxchg.org

In case of multiple NUMA nodes systems, the watermark boosting related code allow kswapd to wake up if one node’s memory is full but other nodes have some free memory. This is for fragmentation handling, but this frequent kswapd work makes some workloads slow. This commit avoids such unnecessary kswapd wakeup.

[PATCH v4 0/1] Add tracepoints around mmap_lock acquisition (Axel Rasmussen)

https://lkml.kernel.org/r/20201020184746.300555-1-axelrasmussen@google.com

Fourth version of the patch adding tracepoints for mmap_lock acquisition. This will help understanding and solving the mmap locking contention problem.

[PATCH 0/2] block layer filter and block device snapshot module (Sergei Shtepa)

https://lkml.kernel.org/r/1603271049-20681-1-git-send-email-sergei.shtepa@veeam.com

This patchset introduce blk-filter, which allows to intercept BIO requests, and blk-snap, which snapshots block device for backup.

Linux 5.10-rc1 (Linus Torvalds)

https://lkml.kernel.org/r/CAHk-=whcRFYSm0jj3Xh3xCyBaxCHA1ZMNO0h_gZso_WZFDUtiQ@mail.gmail.com

Two weeks of the merge window passed, and the 5.10-rc1 has released as a result. Torvalds says this is bigger than he expected. This could be a general upward trend, just a fluke, or due to delayed 5.9 release, Torvalds says. He also picks up the set_fs() removal change from Christoph as most interesting change in this release.

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 2020-10-26 | ~/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 -rc1 releases in the last two years.