Hi Linus, Can you please pull the XFS updates from the tag below? The changes are detailed in the tag message, so I won't go over them again here. The only thing of note is the direct IO fixes that we merged last week after the window opened. Even though a little late, they fix a user reported data corruption and have been pretty well tested. I figured there was not much point waiting another 2 weeks for -rc1 to be released just so I could send them to you.... Note that there will be merge conflicts when you merge this tag. The resolution is pretty straight forward but there's also a direct IO API change that will throw build errors if not fixed during the merge. From Stephen Rothwell's linux-next merge/build process, the DIO API fix that needs to be made is here: http://oss.sgi.com/archives/xfs/2015-04/msg00206.html And the merge resolution is here: http://oss.sgi.com/archives/xfs/2015-04/msg00207.html I would have simple attached a merge commit patch if I could have worked out how to get just the above changes from the merge commit. I can get a combined diff that has lots of noise in it via git show, but I couldn't find a way just to output a simple unified diff. Is there any trick to this, or any specific way you'd prefer that I communicate such issues in future? -Dave. The following changes since commit c517d838eb7d07bbe9507871fab3931deccff539: Linux 4.0-rc1 (2015-02-22 18:21:14 -0800) are available in the git repository at: git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs.git tags/xfs-for-linus-4.1-rc1 for you to fetch changes up to 542c311813d5cb2e6f0dfa9557f41c829b8fb6a0: Merge branch 'xfs-dio-extend-fix' into for-next (2015-04-16 22:13:18 +1000) ---------------------------------------------------------------- xfs: update for 4.1-rc1 This update contains: o RENAME_WHITEOUT support o conversion of per-cpu superblock accounting to use generic counters o new inode mmap lock so that we can lock page faults out of truncate, hole punch and other direct extent manipulation functions to avoid racing mmap writes from causing data corruption o rework of direct IO submission and completion to solve data corruption issue when running concurrent extending DIO writes. Also solves problem of running IO completion transactions in interrupt context during size extending AIO writes. o FALLOC_FL_INSERT_RANGE support for inserting holes into a file via direct extent manipulation to avoid needing to copy data within the file o attribute block header field overflow fix for 64k block size filesystems o Lots of changes to log messaging to be more informative and concise when errors occur. Also prevent a lot of unnecessary log spamming due to cascading failures in error conditions. o lots of cleanups and bug fixes ---------------------------------------------------------------- Brian Foster (3): xfs: pass attr geometry to attr leaf header conversion functions xfs: use larger in-core attr firstused field and detect overflow xfs: kill unnecessary firstused overflow check on attr3 leaf removal Byoungyoung Lee (1): xfs: xfs_mru_cache_insert() should use GFP_NOFS Christoph Hellwig (1): xfs: unlock i_mutex in xfs_break_layouts Dave Chinner (38): xfs: use generic percpu counters for inode counter xfs: use generic percpu counters for free inode counter xfs: use generic percpu counters for free block counter xfs: Remove icsb infrastructure xfs: introduce xfs_mod_frextents xfs: replace xfs_mod_incore_sb_batched xfs: remove xfs_mod_incore_sb API xfs: introduce mmap/truncate lock xfs: use i_mmaplock on read faults xfs: use i_mmaplock on write faults xfs: take i_mmap_lock on extent manipulation operations xfs: xfs_setattr_size no longer races with page faults xfs: lock out page faults from extent swap operations xfs: ensure truncate forces zeroed blocks to disk xfs: inodes are new until the dentry cache is set up xfs: xfs_alloc_fix_minleft can underflow near ENOSPC Merge branch 'xfs-misc-fixes-for-4.1' into for-next Merge branch 'xfs-generic-sb-counters' into for-next Merge branch 'xfs-mmap-lock' into for-next xfs: clean up inode locking for RENAME_WHITEOUT xfs: cleanup xfs_rename error handling xfs: factor out xfs_finish_rename() xfs: make xfs_cross_rename() complete fully xfs: add RENAME_WHITEOUT support Merge branch 'xfs-rename-whiteout' into for-next xfs: remove xfs_bmap_sanity_check() Merge branch 'xfs-misc-fixes-for-4.1-2' into for-next Merge branch 'fallocate-insert-range' into for-next Merge branch 'xfs-misc-fixes-for-4.1-3' into for-next xfs: factor DIO write mapping from get_blocks xfs: move DIO mapping size calculation xfs: DIO needs an ioend for writes xfs: handle DIO overwrite EOF update completion correctly xfs: DIO writes within EOF don't need an ioend xfs: DIO write completion size updates race xfs: direct IO EOF zeroing needs to drain AIO xfs: using generic_file_direct_write() is unnecessary Merge branch 'xfs-dio-extend-fix' into for-next Eric Sandeen (9): xfs: pass mp to XFS_WANT_CORRUPTED_GOTO xfs: pass mp to XFS_WANT_CORRUPTED_RETURN xfs: Ensure we have target_ip for RENAME_EXCHANGE xfs: log unmount events on console xfs: clarify async write failure ratelimit message xfs: cancel failed transaction in xfs_fs_commit_blocks() xfs: remove deprecated mount options xfs: fix NULL pointer dereference in xfs_filestream_lookup_ag() xfs: disallow ro->rw remount on norecovery mount Fabian Frederick (2): xfs: use bool instead of int in xfs_rename() xfs: fix shadow warning in xfs_da3_root_split() Jan Kara (1): xfs: Fix quota type in quota structures when reusing quota file Joe Perches (1): xfs: Fix incorrect positive ENOMEM return Namjae Jeon (2): fs: Add support FALLOC_FL_INSERT_RANGE for fallocate xfs: Add support FALLOC_FL_INSERT_RANGE for fallocate Scott Wood (1): xfs: %pF is only for function pointers Wang Sheng-Hui (1): xfs: remove old and redundant comment in xfs_mount_validate_sb kbuild test robot (1): xfs: xfs_shift_file_space can be static Documentation/filesystems/xfs.txt | 29 +- fs/open.c | 8 +- fs/xfs/libxfs/xfs_alloc.c | 104 +++-- fs/xfs/libxfs/xfs_attr_leaf.c | 150 ++++-- fs/xfs/libxfs/xfs_attr_leaf.h | 6 +- fs/xfs/libxfs/xfs_bmap.c | 554 +++++++++++++++------- fs/xfs/libxfs/xfs_bmap.h | 13 +- fs/xfs/libxfs/xfs_btree.c | 24 +- fs/xfs/libxfs/xfs_da_btree.c | 8 +- fs/xfs/libxfs/xfs_da_format.h | 14 +- fs/xfs/libxfs/xfs_dir2_data.c | 39 +- fs/xfs/libxfs/xfs_format.h | 62 --- fs/xfs/libxfs/xfs_ialloc.c | 48 +- fs/xfs/libxfs/xfs_sb.c | 20 +- fs/xfs/xfs_aops.c | 270 +++++++---- fs/xfs/xfs_attr_inactive.c | 3 +- fs/xfs/xfs_attr_list.c | 9 +- fs/xfs/xfs_bmap_util.c | 164 ++++--- fs/xfs/xfs_bmap_util.h | 2 + fs/xfs/xfs_buf_item.c | 4 +- fs/xfs/xfs_discard.c | 2 +- fs/xfs/xfs_error.c | 2 +- fs/xfs/xfs_error.h | 8 +- fs/xfs/xfs_file.c | 175 +++++-- fs/xfs/xfs_filestream.c | 2 +- fs/xfs/xfs_fsops.c | 20 +- fs/xfs/xfs_icache.c | 4 +- fs/xfs/xfs_inode.c | 554 +++++++++++++--------- fs/xfs/xfs_inode.h | 58 ++- fs/xfs/xfs_ioctl.c | 7 +- fs/xfs/xfs_iomap.c | 3 +- fs/xfs/xfs_iops.c | 127 ++--- fs/xfs/xfs_iops.h | 2 - fs/xfs/xfs_itable.c | 2 +- fs/xfs/xfs_linux.h | 9 - fs/xfs/xfs_log_recover.c | 4 +- fs/xfs/xfs_mount.c | 918 +++++-------------------------------- fs/xfs/xfs_mount.h | 95 +--- fs/xfs/xfs_mru_cache.c | 2 +- fs/xfs/xfs_pnfs.c | 11 +- fs/xfs/xfs_pnfs.h | 5 +- fs/xfs/xfs_qm.c | 18 +- fs/xfs/xfs_super.c | 132 ++++-- fs/xfs/xfs_super.h | 2 + fs/xfs/xfs_symlink.c | 58 ++- fs/xfs/xfs_trace.h | 29 +- fs/xfs/xfs_trans.c | 234 ++++++---- include/linux/falloc.h | 6 + include/uapi/linux/falloc.h | 17 + 49 files changed, 2061 insertions(+), 1976 deletions(-) -- Dave Chinner david@xxxxxxxxxxxxx _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
