OOM killer not nearly agressive enough? https://lkml.kernel.org/r/20200107204412.GA29562@amd
The author if this mail, Pavel updated his userspace to x86-64 and encountered chromium’s memory hogging. He was required to kill the chromium himself, because OOM killer didn’t kill it. He asks whether this means OOM killer is not aggressive enough.
[PATCH 0/4] introduce memory hinting API for external process https://lkml.kernel.org/r/20200110213433.94739-1-minchan@kernel.org
Minchan has previously posted the patchset introducing the MADV_COLD, MADV_PAGEOUT and external madvise() able code change.
[PATCH v16 QEMU 4/3 RFC] memory: Add support for MADV_FREE as mechanism to lazy discard pages https://lkml.kernel.org/r/20200103212339.29849.99817.stgit@localhost.localdomain
This RFC patch modifies QEMU to able to use MADV_FREE for discard of pages to avoid the page zeroing cost.
Linux 5.5-rc5 https://lkml/kernel.org/r/CAHk-=whwmJQwwMD06VfPapDmc52LFQhG3HP6VyVLXVA=nYi_ow@mail.gmail.com
Time files, rc5 released. Last week was also another calm week due to the new year holidays, and thus this release is small. Linus expects it will be changed in this upcoming weeks, or 5.
!! NOTE !!
This post has migrated to https://damonitor.github.io/posts/damon. This out-dated post will be removed soon.
With increasingly data-intensive workloads and limited DRAM capacity, optimal memory management based on dynamic access patterns is becoming increasingly important. Such mechanisms are only possible if accurate and efficient dynamic access pattern monitoring is available.
DAMON is a Data Access MONitoring framework subsystem for the Linux kernel developed for such memory management. It is designed with some key mechanism (refer to Design for the detail) that make it
[PATCH V11 RESEND] mm/debug: Add tests validating architecture page table helpers https://lkml.kernel.org/r/1577162322-30303-1-git-send-email-anshuman.khandual@arm.com
This patchset adds tests for the page table helpers. Those tests covers basic page table entry transformations including old, young, dirty, clean, write, write protect etc. The tests are not using the kselftest or kunit, but implemented as a kernel module with new file, mm/debug_vm_pgtable.c.
Linux 5.5-rc4 https://lkml.kernel.org/r/CAHk-=wjq2b9GkLzpaDE-Xryu5c0zMM72BqkJKeZVsX+4ymH6aA@mail.gmail.com
Due to the holidays, last week was unsurprisingly quiet. Torvalds says it was hardly even worth making an rc release.
PROBLEM: Large (256MB) CMA allocation fails from time to time when using Intel VT-d https://lkml.kernel.org/r/5e75d900cb50457784ee1d2050b511ec@MAILER13.tech.ceia
The author reports a CMA allocation failure that occurs time to time if the system is using Intel VT-d.
[PATCHv2 0/3] Allow ZRAM to use any zpool-compatible backend https://lkml.kernel.org/r/20191219151928.ad4ccf732b64b7f8a26116db@gmail.com
ZRAM is an in-memory block device that compresses the content. Currently, ZRAM is used only with zsmalloc. However, the zsmalloc is not an optimal selection in some environment.
[PATCHSET 0/5] Support for RWF_UNCACHED https://lkml.kernel.org/r/20191210162454.8608-1-axboe@kernel.dk
As the buffered IO uses the page cache, an intensive buffered IO workload using very huge file can cause memory pressure by inflating the page cache. Changing the workload to do direct IO can solve this problem, but the modification is not so simple. Jens Axboe thus implemented RWF_UNCACHED flag, which skips use of the page cache if the given IO target is not in the page cache already.
Splitting the mmap_sem https://lkml.kernel.org/r/20191203222147.GV20752@bombadil.infradead.org
The mmap_sem is a big scalability bottleneck and there are several projects for solving it. In this mail, Matthew Wilcox suggests his plan for his solution and asks whether a exceptional case in his solution is acceptable to other people.
Linux 5.5-rc1 https://lkml.kernel.org/r/CAHk-=wiEJK=yo9vEhX_4b4ROvCqUA_rjK7g996h-5MbfOMeDrw@mail.gmail.com
Linux 5.5 merge window closed and the -rc1 has released. It was a tiny bit larger release compared to recent several releases, but fairly regular overall.
[PATCH RFC v3 00/36] Add KernelMemorySanitizer infrastructure https://lkml.kernel.org/r/20191122112621.204798-1-glider@google.com
Google’s third KMSAN RFC patchset that detects errors caused by accessing uninitialized memory areas. Like other SAN series, it has been used in Google syzkaller and found over 200 bugs in the last two years.
LSF/MM/BPF: 2020: Call for Proposals https://lkml.kernel.org/r/20191122172502.vffyfxlqejthjib6@macbook-pro-91.dhcp.thefacebook.com
LSF/MM’s CFP for the next year was released. From next year, BPF sessions will be added to LSF/MM, so the name has changed to LSF/MM/BPF.
[PATCH 00/13] Finish off [smp_]read_barrier_depends() https://lkml.kernel.org/r/20191108170120.22331-1-will@kernel.org
The memory model element which is depends only on Alpha, smp_read_barrier_depends() is built into READ_ONCE(), so that you don’t have to worry about Alpha in non-alpha-specific kernel code. Nevertheless, the author of this patchset concerned about the continuing uploads of the code that cares it. This patchset removes the description of the feature from all documentation and comments in the kernel code.
[PATCH v4 00/10] Add Kernel Concurrency Sanitizer (KCSAN) https://lkml.
Linux 5.4-rc7 https://lkml.kernel.org/r/CAHk-=wivmGEOTha5XkRHgH6VwfiroiN+PFVMNK3B0r-d0eFLPQ@mail.gmail.com
일곱번째 rc 버전이 릴리즈 되었습니다. 지난 rc6 때도 변경량이 생각보다 많아 토발즈가 고민했는데, 이번주도 토발즈가 원했던 것보단 변경량이 많았다고 하는군요. 다음주에 rc8 을 낼지 5.4 정식 버전을 낼지는 좀 더 고민해 봐야겠다고 합니다.