Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu,  8 Jun 2023 17:51:52 -0700
Mime-Version: 1.0
Message-ID: <20230609005158.2421285-1-surenb@google.com>
Subject: [PATCH v2 0/6] Per-vma lock support for swap and userfaults
From:   Suren Baghdasaryan <surenb@google.com>
To:     akpm@linux-foundation.org
Cc:     willy@infradead.org, hannes@cmpxchg.org, mhocko@suse.com,
        josef@toxicpanda.com, jack@suse.cz, ldufour@linux.ibm.com,
        laurent.dufour@fr.ibm.com, michel@lespinasse.org,
        liam.howlett@oracle.com, jglisse@google.com, vbabka@suse.cz,
        minchan@google.com, dave@stgolabs.net, punit.agrawal@bytedance.com,
        lstoakes@gmail.com, hdanton@sina.com, apopple@nvidia.com,
        peterx@redhat.com, ying.huang@intel.com, david@redhat.com,
        yuzhao@google.com, dhowells@redhat.com, hughd@google.com,
        viro@zeniv.linux.org.uk, brauner@kernel.org,
        pasha.tatashin@soleen.com, surenb@google.com, linux-mm@kvack.org,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        kernel-team@android.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

When per-vma locks were introduced in [1] several types of page faults
would still fall back to mmap_lock to keep the patchset simple. Among them
are swap and userfault pages. The main reason for skipping those cases was
the fact that mmap_lock could be dropped while handling these faults and
that required additional logic to be implemented.
Implement the mechanism to allow per-vma locks to be dropped for these
cases. When that happens handle_mm_fault returns new VM_FAULT_VMA_UNLOCKED
vm_fault_reason bit along with VM_FAULT_RETRY to indicate that VMA lock
was dropped. Naturally, once VMA lock is dropped that VMA should be
assumed unstable and can't be used.

Changes since v1 posted at [2]
- New patch 1/6 to remove do_poll parameter from read_swap_cache_async(),
per Huang Ying
- New patch 3/6 to separate VM_FAULT_COMPLETED addition,
per Alistair Popple
- Added comment for VM_FAULT_VMA_UNLOCKED in 4/6, per Alistair Popple
- New patch 6/6 to handle userfaults under VMA lock

Note: I tried implementing Matthew's suggestion in [3] to add vmf_end_read
but that gets quite messy since it would require changing code for every
architecture when we change handle_mm_fault interface.

Note: patch 4/6 will cause a trivial merge conflict in arch/arm64/mm/fault.c
when applied over mm-unstable branch due to a patch from ARM64 tree [4]
which is missing in mm-unstable.

[1] https://lore.kernel.org/all/20230227173632.3292573-1-surenb@google.com/
[2] https://lore.kernel.org/all/20230501175025.36233-1-surenb@google.com/
[3] https://lore.kernel.org/all/ZFEeHqzBJ6iOsRN+@casper.infradead.org/
[4] https://lore.kernel.org/all/20230524131305.2808-1-jszhang@kernel.org/

Suren Baghdasaryan (6):
  swap: remove remnants of polling from read_swap_cache_async
  mm: handle swap page faults under VMA lock if page is uncontended
  mm: add missing VM_FAULT_RESULT_TRACE name for VM_FAULT_COMPLETED
  mm: drop VMA lock before waiting for migration
  mm: implement folio wait under VMA lock
  mm: handle userfaults under VMA lock

 arch/arm64/mm/fault.c    |  3 ++-
 arch/powerpc/mm/fault.c  |  3 ++-
 arch/s390/mm/fault.c     |  3 ++-
 arch/x86/mm/fault.c      |  3 ++-
 fs/userfaultfd.c         | 42 ++++++++++++++++++-----------------
 include/linux/mm_types.h |  7 +++++-
 include/linux/pagemap.h  | 14 ++++++++----
 mm/filemap.c             | 37 +++++++++++++++++++------------
 mm/madvise.c             |  4 ++--
 mm/memory.c              | 48 ++++++++++++++++++++++------------------
 mm/swap.h                |  1 -
 mm/swap_state.c          | 12 +++++-----
 12 files changed, 103 insertions(+), 74 deletions(-)

-- 
2.41.0.162.gfafddb0af9-goog