mbox series

[mm-unstable,v4,0/7] mm/mglru: performance optimizations

Message ID 20241231043538.4075764-1-yuzhao@google.com (mailing list archive)
Headers show
Series mm/mglru: performance optimizations | expand

Message

Yu Zhao Dec. 31, 2024, 4:35 a.m. UTC 
This series improves performance for some previously reported test
cases. Most of the code changes gathered here has been floating on the
mailing list [1][2]. They are now properly organized and have gone
through various benchmarks on client and server devices, including
Android, FIO, memcached, multiple VMs and MongoDB.

In addition to the syzbot regressions fixed in v2 [3] and v3 [4], this
version fixes two more regressions: one reported by Oliver Sang [5]
and the other by Barry Song.

[1] https://lore.kernel.org/CAOUHufahuWcKf5f1Sg3emnqX+cODuR=2TQo7T4Gr-QYLujn4RA@mail.gmail.com/
[2] https://lore.kernel.org/CAOUHufawNerxqLm7L9Yywp3HJFiYVrYO26ePUb1jH-qxNGWzyA@mail.gmail.com/
[3] https://lore.kernel.org/67294349.050a0220.701a.0010.GAE@google.com/
[4] https://lore.kernel.org/67549eca.050a0220.2477f.001b.GAE@google.com/
[5] https://lore.kernel.org/202412231601.f1eb8f84-lkp@intel.com/

Yu Zhao (7):
  mm/mglru: clean up workingset
  mm/mglru: optimize deactivation
  mm/mglru: rework aging feedback
  mm/mglru: rework type selection
  mm/mglru: rework refault detection
  mm/mglru: rework workingset protection
  mm/mglru: fix PTE-mapped large folios

 include/linux/mm_inline.h |  88 ++++---
 include/linux/mmzone.h    |  99 +++++---
 mm/swap.c                 |  70 ++++--
 mm/vmscan.c               | 515 +++++++++++++++++++-------------------
 mm/workingset.c           |  67 +++--
 5 files changed, 445 insertions(+), 394 deletions(-)

Comments

Andrew Morton Jan. 3, 2025, 12:03 a.m. UTC | #1 
On Mon, 30 Dec 2024 21:35:31 -0700 Yu Zhao <yuzhao@google.com> wrote:

> This series improves performance for some previously reported test
> cases. Most of the code changes gathered here has been floating on the
> mailing list [1][2]. They are now properly organized and have gone
> through various benchmarks on client and server devices, including
> Android, FIO, memcached, multiple VMs and MongoDB.

This series has significant conflicts with the patch "mm: vmscan: retry
folios written back while isolated for traditional LRU".  It appears
that "mm: vmscan: retry folios written back while isolated for
traditional LRU" is due for an update and that more discussion is
needed, so I shall drop this version of "mm: vmscan: retry folios
written back while isolated for traditional LRU" from mm-unstable.
chenridong Jan. 4, 2025, 12:59 a.m. UTC | #2 
On 2025/1/3 8:03, Andrew Morton wrote:
> On Mon, 30 Dec 2024 21:35:31 -0700 Yu Zhao <yuzhao@google.com> wrote:
> 
>> This series improves performance for some previously reported test
>> cases. Most of the code changes gathered here has been floating on the
>> mailing list [1][2]. They are now properly organized and have gone
>> through various benchmarks on client and server devices, including
>> Android, FIO, memcached, multiple VMs and MongoDB.
> 
> This series has significant conflicts with the patch "mm: vmscan: retry
> folios written back while isolated for traditional LRU".  It appears
> that "mm: vmscan: retry folios written back while isolated for
> traditional LRU" is due for an update and that more discussion is
> needed, so I shall drop this version of "mm: vmscan: retry folios
> written back while isolated for traditional LRU" from mm-unstable.

Hi, Andrew and Yu, does this mean I should resend a new version that
fixes the conflict and updates the message?

Best regards,
Ridong
Andrew Morton Jan. 4, 2025, 2:21 a.m. UTC | #3 
On Sat, 4 Jan 2025 08:59:08 +0800 chenridong <chenridong@huawei.com> wrote:

> 
> 
> On 2025/1/3 8:03, Andrew Morton wrote:
> > On Mon, 30 Dec 2024 21:35:31 -0700 Yu Zhao <yuzhao@google.com> wrote:
> > 
> >> This series improves performance for some previously reported test
> >> cases. Most of the code changes gathered here has been floating on the
> >> mailing list [1][2]. They are now properly organized and have gone
> >> through various benchmarks on client and server devices, including
> >> Android, FIO, memcached, multiple VMs and MongoDB.
> > 
> > This series has significant conflicts with the patch "mm: vmscan: retry
> > folios written back while isolated for traditional LRU".  It appears
> > that "mm: vmscan: retry folios written back while isolated for
> > traditional LRU" is due for an update and that more discussion is
> > needed, so I shall drop this version of "mm: vmscan: retry folios
> > written back while isolated for traditional LRU" from mm-unstable.
> 
> Hi, Andrew and Yu, does this mean I should resend a new version that
> fixes the conflict and updates the message?

Yes please.  Against
https://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm.git/'s
mm-unstable branch would be ideal.

Please already double-check that all review comments have been
addressed in an appropriate fashion.