[GIT PULL] gfs2: 4.19 merge

[Andreas Gruenbacher]

Hi Linus,

please consider pulling the following changes for gfs2:

  1. Iomap support for buffered writes and for direct I/O.
  2. Two patches that reduce the size of struct gfs2_inode.
  3. Lots of fixes and cleanups.

The following merge commit is non-empty on purpuse:

  Merge branch 'iomap-4.19-merge' into linux-gfs2/for-next

There we set the IOMAP_F_BUFFER_HEAD flag: this flag is introduced by commit
"iomap: add initial support for writes without buffer heads"; setting it after
the merge commit would break bisecting.

Thanks,
Andreas

-- 

The following changes since commit 806a1477b10a153cd01ee7ccba8ca2492df3e0b2:

  iomap: add inline data support to iomap_readpage_actor (2018-07-03 09:07:47 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2.git gfs2-4.19.fixes

for you to fetch changes up to f5580d0f8bf60993a5fbc73ee04678070ffbba57:

  gfs2: eliminate update_rgrp_lvb_unlinked (2018-08-08 10:34:39 -0500)

----------------------------------------------------------------
Andreas Gruenbacher (16):
      gfs2: Don't withdraw under a spin lock
      gfs2: Minor clarification to __gfs2_punch_hole
      gfs2: Further iomap cleanups
      gfs2: iomap buffered write support
      gfs2: gfs2_extent_length cleanup
      gfs2: iomap direct I/O support
      gfs2: Remove gfs2_write_{begin,end}
      gfs2: Stop messing with ip->i_rgd in the rlist code
      gfs2: Eliminate redundant ip->i_rgd
      Merge branch 'iomap-write' into linux-gfs2/for-next
      gfs2: fallocate_chunk: Always initialize struct iomap
      Merge branch 'iomap-4.19-merge' into linux-gfs2/for-next
      gfs2: Use iomap for stuffed direct I/O reads
      gfs2: use iomap_readpage for blocksize == PAGE_SIZE
      gfs2: Special-case rindex for gfs2_grow
      gfs2: Get rid of gfs2_ea_strlen

Arnd Bergmann (1):
      gfs2: call ktime_get_coarse_real_ts64() directly

Bob Peterson (7):
      gfs2: eliminate rs_inum and reduce the size of gfs2 inodes
      gfs2: Don't reject a supposedly full bitmap if we have blocks reserved
      GFS2: Fix recovery issues for spectators
      GFS2: rgrp free blocks used incorrectly
      gfs2: cleanup: call gfs2_rgrp_ondisk2lvb from gfs2_rgrp_out
      gfs2: Fix gfs2_testbit to use clone bitmaps
      gfs2: eliminate update_rgrp_lvb_unlinked

Chengguang Xu (1):
      gfs2: using posix_acl_xattr_size instead of posix_acl_to_xattr

Colin Ian King (1):
      gfs2: remove redundant variable 'moved'

Souptick Joarder (1):
      fs: gfs2: Adding new return type vm_fault_t

 fs/gfs2/acl.c        |   6 +-
 fs/gfs2/aops.c       | 337 +++--------------------------------------
 fs/gfs2/aops.h       |  19 +++
 fs/gfs2/bmap.c       | 414 +++++++++++++++++++++++++++++++++++++++++----------
 fs/gfs2/dir.c        |   4 +-
 fs/gfs2/file.c       | 161 ++++++++++++++++++--
 fs/gfs2/incore.h     |  23 ++-
 fs/gfs2/lock_dlm.c   |  20 ++-
 fs/gfs2/log.c        |  28 +++-
 fs/gfs2/lops.c       |   2 +-
 fs/gfs2/meta_io.c    |   4 +-
 fs/gfs2/recovery.c   |   7 +-
 fs/gfs2/rgrp.c       | 169 ++++++++++++---------
 fs/gfs2/super.c      |   1 -
 fs/gfs2/sys.c        |  11 +-
 fs/gfs2/trace_gfs2.h |   3 +-
 fs/gfs2/trans.h      |   6 +-
 fs/gfs2/util.c       |  38 ++---
 fs/gfs2/util.h       |  10 +-
 fs/gfs2/xattr.c      |  59 +++-----
 20 files changed, 752 insertions(+), 570 deletions(-)
 create mode 100644 fs/gfs2/aops.h