linux-kernel - [PATCH v4 0/8] gfs2: Fix mmap + page fault deadlocks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20210724193449.361667-1-agruenba@redhat.com>
Date:   Sat, 24 Jul 2021 21:34:41 +0200
From:   Andreas Gruenbacher <agruenba@...hat.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Alexander Viro <viro@...iv.linux.org.uk>,
        Christoph Hellwig <hch@...radead.org>,
        "Darrick J. Wong" <djwong@...nel.org>
Cc:     Jan Kara <jack@...e.cz>, Matthew Wilcox <willy@...radead.org>,
        cluster-devel@...hat.com, linux-fsdevel@...r.kernel.org,
        linux-kernel@...r.kernel.org, ocfs2-devel@....oracle.com,
        Andreas Gruenbacher <agruenba@...hat.com>
Subject: [PATCH v4 0/8] gfs2: Fix mmap + page fault deadlocks

Hi Linus et al.,

here's another update of this patch queue:

 * Finally fix the typos Linus has pointed out twice already.

 * Turn the previous fault_in_iov_iter helper that was used for reads
   and writes into iov_iter_fault_in_writeable per Al's suggestion.
   Use the existing iov_iter_fault_in_readable for writes.

 * Add a done_before argument and an IOMAP_DIO_FAULT_RETRY flag to
   iomap_dio_rw to allow iomap_dio_rw to return partial results and
   resume with the rest of a request.  This allows iomap_dio_rw to be
   used with page faults disabled without having to repeat any I/O.

 * Adjust the gfs2 patches accordingly.

With that, the two iov_ter patches and the three iomap patches should
hopefully be ready for mainline.

There's one remaining issue on the gfs2 side: during read requests, when
a writer now comes in in the middle of a read request, the read request
can currently return a result that never existed on disk.  So we need
to ensure that we only resume read requests when we know that no writer
got in the way, and retry the entire request otherwise.  It should be
relatively easy to add a mechanism to detect when a glock is "lost";
this won't affect the vfs or iomap patches.

Thanks a lot,
Andreas

Andreas Gruenbacher (8):
  iov_iter: Introduce iov_iter_fault_in_writeable helper
  gfs2: Add wrapper for iomap_file_buffered_write
  gfs2: Fix mmap + page fault deadlocks for buffered I/O
  iomap: Fix iomap_dio_rw return value for user copies
  iomap: Add done_before argument to iomap_dio_rw
  iomap: Support restarting direct I/O requests after user copy failures
  iov_iter: Introduce noio flag to disable page faults
  gfs2: Fix mmap + page fault deadlocks for direct I/O

 fs/btrfs/file.c       |  5 ++-
 fs/ext4/file.c        |  5 ++-
 fs/gfs2/file.c        | 95 +++++++++++++++++++++++++++++++++++++++----
 fs/iomap/direct-io.c  | 29 ++++++++++---
 fs/xfs/xfs_file.c     |  6 +--
 fs/zonefs/super.c     |  4 +-
 include/linux/iomap.h | 11 ++++-
 include/linux/mm.h    |  3 ++
 include/linux/uio.h   |  2 +
 lib/iov_iter.c        | 60 ++++++++++++++++++++++++---
 mm/gup.c              | 57 ++++++++++++++++++++++++++
 11 files changed, 246 insertions(+), 31 deletions(-)

-- 
2.26.3