On Thu, 21 Mar 2019 at 02:50, Mikhail Gavrilov
<mikhail.v.gavrilov@gmail.com> wrote:
>
> On Sat, 16 Mar 2019 at 01:59, Daniel Jordan <daniel.m.jordan@oracle.com> wrote:
> >
> > This is new code, from e332f741a8dd1 ("mm, compaction: be selective about what
> > pageblocks to clear skip hints"), so I added some folks.
> >
> > Can you show
> > $LINUX/scripts/faddr2line path/to/vmlinux __reset_isolation_pfn+0x244
> > ?
>
> $ /usr/src/kernels/`uname -r`/scripts/faddr2line
> /lib/debug/lib/modules/`uname -r`/vmlinux __reset_isolation_pfn+0x244
> __reset_isolation_pfn+0x244/0x2b0:
> page_to_nid at include/linux/mm.h:1021
> (inlined by) page_zone at include/linux/mm.h:1163
> (inlined by) __reset_isolation_pfn at mm/compaction.c:250
>
> It was not easy, but I completed just now kernel bisecting and see
> that you right.
> First bad commit is e332f741a8dd1
>
> $ git bisect log
> git bisect start
> # good: [cd2a3bf02625ffad02a6b9f7df758ee36cf12769] Merge tag
> 'leds-for-5.1-rc1' of
> git://git.kernel.org/pub/scm/linux/kernel/git/j.anaszewski/linux-leds
> git bisect good cd2a3bf02625ffad02a6b9f7df758ee36cf12769
> # bad: [610cd4eadec4f97acd25d3108b0e50d1362b3319] Merge branch
> 'x86-uv-for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> git bisect bad 610cd4eadec4f97acd25d3108b0e50d1362b3319
> # good: [203b6609e0ede49eb0b97008b1150c69e9d2ffd3] Merge branch
> 'perf-core-for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
> git bisect good 203b6609e0ede49eb0b97008b1150c69e9d2ffd3
> # bad: [da2577fe63f865cd9dc785a42c29c0071f567a35] Merge tag
> 'sound-5.1-rc1' of
> git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
> git bisect bad da2577fe63f865cd9dc785a42c29c0071f567a35
> # good: [fb686ad25be0343a9dab23acff674d0cb84bb516] Merge tag
> 'armsoc-defconfig' of
> git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
> git bisect good fb686ad25be0343a9dab23acff674d0cb84bb516
> # good: [70395a96bd882d8dba669f99b5cec0008690accd] Merge tag
> 'asoc-v5.1-2' of
> https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into
> for-next
> git bisect good 70395a96bd882d8dba669f99b5cec0008690accd
> # bad: [8dcd175bc3d50b78413c56d5b17d4bddd77412ef] Merge branch 'akpm'
> (patches from Andrew)
> git bisect bad 8dcd175bc3d50b78413c56d5b17d4bddd77412ef
> # bad: [7f18825174203526a47c127c12a50f897ee0b511] powerpc/mm/iommu:
> allow large IOMMU page size only for hugetlb backing
> git bisect bad 7f18825174203526a47c127c12a50f897ee0b511
> # good: [566e54e113eb2b669f9300db2c2df400cbb06646] mm, compaction:
> remove last_migrated_pfn from compact_control
> git bisect good 566e54e113eb2b669f9300db2c2df400cbb06646
> # bad: [d9f7979c92f7b34469c1ca5d1f3add6681fd567c] mm: no need to check
> return value of debugfs_create functions
> git bisect bad d9f7979c92f7b34469c1ca5d1f3add6681fd567c
> # good: [cb810ad294d3c3a454e51b12fbb483bbb7096b98] mm, compaction:
> rework compact_should_abort as compact_check_resched
> git bisect good cb810ad294d3c3a454e51b12fbb483bbb7096b98
> # bad: [147e1a97c4a0bdd43f55a582a9416bb9092563a9] fs: kernfs: add poll
> file operation
> git bisect bad 147e1a97c4a0bdd43f55a582a9416bb9092563a9
> # good: [dbe2d4e4f12e07c6a2215e3603a5f77056323081] mm, compaction:
> round-robin the order while searching the free lists for a target
> git bisect good dbe2d4e4f12e07c6a2215e3603a5f77056323081
> # bad: [e332f741a8dd1ec9a6dc8aa997296ecbfe64323e] mm, compaction: be
> selective about what pageblocks to clear skip hints
> git bisect bad e332f741a8dd1ec9a6dc8aa997296ecbfe64323e
> # good: [4fca9730c51d51f643f2a3f8f10ebd718349c80f] mm, compaction:
> sample pageblocks for free pages
> git bisect good 4fca9730c51d51f643f2a3f8f10ebd718349c80f
> # first bad commit: [e332f741a8dd1ec9a6dc8aa997296ecbfe64323e] mm,
> compaction: be selective about what pageblocks to clear skip hints
>
> Also I see that two patches already proposed for fixing this issue.
> [1] https://patchwork.kernel.org/patch/10862267/
> [2] https://patchwork.kernel.org/patch/10862519/
>
> If I understand correctly, it is enough to apply only the second patch [2].
>

I am right now tested the patch [1] and can said that unfortunately it
not fix my issue.
[1] https://patchwork.kernel.org/patch/10862519/

I am attached full kernel log here.

How issue reproduced:
1) Application with heavy I/O activity eat memory for disk cache. (For
example steam client downloads heavy game 50Gb)
2) And when starts using swap this kernel panic is happened.

My system specs:
RAM: 32GB
Swap: 64GB


--
Best Regards,
Mike Gavrilov.