From: Suren Baghdasaryan <surenb@google.com>
To: akpm@linux-foundation.org
Cc: willy@infradead.org, liam.howlett@oracle.com,
lorenzo.stoakes@oracle.com, mhocko@suse.com, vbabka@suse.cz,
hannes@cmpxchg.org, mjguzik@gmail.com, oliver.sang@intel.com,
mgorman@techsingularity.net, david@redhat.com,
peterx@redhat.com, oleg@redhat.com, dave@stgolabs.net,
paulmck@kernel.org, brauner@kernel.org, dhowells@redhat.com,
hdanton@sina.com, hughd@google.com, minchan@google.com,
jannh@google.com, shakeel.butt@linux.dev,
souravpanda@google.com, pasha.tatashin@soleen.com,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
kernel-team@android.com, surenb@google.com
Subject: [PATCH v2 0/5] move per-vma lock into vm_area_struct
Date: Tue, 12 Nov 2024 11:46:30 -0800 [thread overview]
Message-ID: <20241112194635.444146-1-surenb@google.com> (raw)
Back when per-vma locks were introduces, vm_lock was moved out of
vm_area_struct in [1] because of the performance regression caused by
false cacheline sharing. Recent investigation [2] revealed that the
regressions is limited to a rather old Broadwell microarchitecture and
even there it can be mitigated by disabling adjacent cacheline
prefetching, see [3].
This patchset moves vm_lock back into vm_area_struct, aligning it at the
cacheline boundary and changing the cache to be cache-aligned as well.
This causes VMA memory consumption to grow from 160 (vm_area_struct) + 40
(vm_lock) bytes to 256 bytes:
slabinfo before:
<name> ... <objsize> <objperslab> <pagesperslab> : ...
vma_lock ... 40 102 1 : ...
vm_area_struct ... 160 51 2 : ...
slabinfo after moving vm_lock:
<name> ... <objsize> <objperslab> <pagesperslab> : ...
vm_area_struct ... 256 32 2 : ...
Aggregate VMA memory consumption per 1000 VMAs grows from 50 to 64 pages,
which is 5.5MB per 100000 VMAs. This overhead will be addressed in a
separate patchset by replacing rw_semaphore in vma_lock's implementation
with a different type of lock.
Moving vm_lock into vm_area_struct lets us simplify vm_area_free() path,
which in turn allows us to use SLAB_TYPESAFE_BY_RCU for vm_area_struct
cache. This should facilitate vm_area_struct reuse and will minimize the
number of call_rcu() calls.
Suren Baghdasaryan (5):
mm: introduce vma_start_read_locked{_nested} helpers
mm: move per-vma lock into vm_area_struct
mm: mark vma as detached until it's added into vma tree
mm: make vma cache SLAB_TYPESAFE_BY_RCU
docs/mm: document latest changes to vm_lock
Documentation/mm/process_addrs.rst | 10 +++--
include/linux/mm.h | 54 +++++++++++++++++-----
include/linux/mm_types.h | 16 ++++---
include/linux/slab.h | 6 ---
kernel/fork.c | 72 +++++++-----------------------
mm/memory.c | 2 +-
mm/mmap.c | 2 +
mm/nommu.c | 2 +
mm/userfaultfd.c | 14 +++---
mm/vma.c | 3 ++
tools/testing/vma/vma_internal.h | 3 +-
11 files changed, 92 insertions(+), 92 deletions(-)
base-commit: 931086f2a88086319afb57cd3925607e8cda0a9f
--
2.47.0.277.g8800431eea-goog
next reply other threads:[~2024-11-12 19:46 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-12 19:46 Suren Baghdasaryan [this message]
2024-11-12 19:46 ` [PATCH v2 1/5] mm: introduce vma_start_read_locked{_nested} helpers Suren Baghdasaryan
2024-11-13 14:10 ` Lorenzo Stoakes
2024-11-13 15:30 ` Suren Baghdasaryan
2024-11-12 19:46 ` [PATCH v2 2/5] mm: move per-vma lock into vm_area_struct Suren Baghdasaryan
2024-11-13 14:28 ` Lorenzo Stoakes
2024-11-13 14:45 ` Vlastimil Babka
2024-11-13 14:58 ` Lorenzo Stoakes
2024-11-13 15:09 ` Vlastimil Babka
2024-11-13 14:53 ` Mateusz Guzik
2024-11-13 14:59 ` Lorenzo Stoakes
2024-11-13 15:01 ` Lorenzo Stoakes
2024-11-13 15:45 ` Suren Baghdasaryan
2024-11-13 15:42 ` Suren Baghdasaryan
2024-11-12 19:46 ` [PATCH v2 3/5] mm: mark vma as detached until it's added into vma tree Suren Baghdasaryan
2024-11-13 14:43 ` Lorenzo Stoakes
2024-11-13 15:37 ` Suren Baghdasaryan
2024-11-12 19:46 ` [PATCH v2 4/5] mm: make vma cache SLAB_TYPESAFE_BY_RCU Suren Baghdasaryan
2024-11-13 2:57 ` Suren Baghdasaryan
2024-11-13 5:08 ` Hugh Dickins
2024-11-13 6:03 ` Suren Baghdasaryan
2024-11-13 6:52 ` Hugh Dickins
2024-11-13 8:19 ` Suren Baghdasaryan
2024-11-13 8:58 ` Vlastimil Babka
2024-11-13 12:38 ` Liam R. Howlett
2024-11-13 13:57 ` Matthew Wilcox
2024-11-13 15:22 ` Liam R. Howlett
2024-11-13 15:25 ` Suren Baghdasaryan
2024-11-13 15:29 ` Liam R. Howlett
2024-11-13 15:47 ` Suren Baghdasaryan
2024-11-13 19:05 ` Suren Baghdasaryan
2024-11-14 16:18 ` Suren Baghdasaryan
2024-11-14 16:21 ` Vlastimil Babka
2024-11-13 16:44 ` Jann Horn
2024-11-13 20:59 ` Matthew Wilcox
2024-11-13 21:23 ` Jann Horn
2024-11-12 19:46 ` [PATCH v2 5/5] docs/mm: document latest changes to vm_lock Suren Baghdasaryan
2024-11-12 19:51 ` Suren Baghdasaryan
2024-11-13 14:46 ` Lorenzo Stoakes
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20241112194635.444146-1-surenb@google.com \
--to=surenb@google.com \
--cc=akpm@linux-foundation.org \
--cc=brauner@kernel.org \
--cc=dave@stgolabs.net \
--cc=david@redhat.com \
--cc=dhowells@redhat.com \
--cc=hannes@cmpxchg.org \
--cc=hdanton@sina.com \
--cc=hughd@google.com \
--cc=jannh@google.com \
--cc=kernel-team@android.com \
--cc=liam.howlett@oracle.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=lorenzo.stoakes@oracle.com \
--cc=mgorman@techsingularity.net \
--cc=mhocko@suse.com \
--cc=minchan@google.com \
--cc=mjguzik@gmail.com \
--cc=oleg@redhat.com \
--cc=oliver.sang@intel.com \
--cc=pasha.tatashin@soleen.com \
--cc=paulmck@kernel.org \
--cc=peterx@redhat.com \
--cc=shakeel.butt@linux.dev \
--cc=souravpanda@google.com \
--cc=vbabka@suse.cz \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox