LKML News v5.6-rc5

[PATCH v1 00/11] virtio-mem: paravirtualized memory

https://lkml.kernel.org/r/20200302134941.315212-1-david@redhat.com

At KVM Forum 2018, virtio-mem’s basic idea has presented and now this patchset introduces it. This provides a flexible, cross-architecture memory hot(un)plug solution.

[PATCH v7 0/7] introduce memory hinting API for external process

https://lkml.kernel.org/r/20200302193630.68771-1-minchan@kernel.org

Finally, this patchset has accepted by Andrew Morton.

[PATCH v2] mm: Proactive compaction

https://lkml.kernel.org/r/20200302213343.2712-1-nigupta@nvidia.com

This patchset makes kernel to do compaction proactively instead of THP allocation failures, to minimize the latency spike which comes from the compaction. It significantly reduces THP allocation latency under memory fragmentation. 95th percentile of the latency has reduced from 33,799 to 429!

[RFC v4 0/7] Implement Data Access Monitoring-based Memory Operation Schemes

https://lkml.kernel.org/r/20200303121406.20954-1-sjpark@amazon.com

My fourth RFC for DAMON-based Operation Schemes. This time, I updated the evaluation results of ETHP scheme with more reliable data.

TL;DR: ‘ethp’ removes 97.61% of ‘thp’ memory space overhead while preserving 25.40% (up to 88.36%) of ‘thp’ performance improvement in total.

[LSFMMBPF TOPIC] Killing LSFMMBPF

https://lkml.kernel.org/r/b506a373-c127-b92e-9824-16e8267fc910@toxicpanda.com

Another LSF/MM/BPF topic proposal. Author suggests killing LSF/MM/BPF. The main points of problems the author feels are LSF/MM/BPF is invitation only but there are too many people need to be involved so that cannot fully invite every necessary people. The author proposing LSF/MM to be somewhat like LPC instead.

[LSFMMBPF TOPIC] LSFMMBPF 2020 COVID-19 status update

https://lkml.kernel.org/r/20200307185420.GG2236@42.do-not-panic.com

COVID-19 is affecting so many conferences. It seems LSF/MM/BPF also not an exception. The committees are trying to figure out what is the right decision for it.

Linux 5.6-rc5

https://lkml.kernel.org/r/CAHk-=wjie=L9=jO-VKR5u7+5Ekis-VR5m=zUChR9djhAuAN55g@mail.gmail.com

Torvalds made this release from 28,000 feet, mainly due to flight cancellations and travel trouble. Maybe due to the COVID-19?

The release itself shows no worrysome things, but slightly bigger than Torvalds’ preferred -rc5 size.