Message-ID: <20161012164707.GF5619@birch.djwong.org>
Date:   Wed, 12 Oct 2016 09:47:07 -0700
From:   "Darrick J. Wong" <darrick.wong@...cle.com>
To:     Dave Chinner <david@...morbit.com>
Cc:     torvalds@...ux-foundation.org, akpm@...ux-foundation.org,
        linux-kernel@...r.kernel.org, linux-xfs@...r.kernel.org
Subject: Re: [GIT PULL] xfs: shared data extents support for 4.9-rc1

On Wed, Oct 12, 2016 at 11:18:49PM +1100, Dave Chinner wrote:
> Hi Linus,
> 
> This is the second part of the XFS updates for this merge cycle.
> This pullreq contains the new shared data extents feature for XFS,
> and can be found at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs.git tags/xfs-reflink-for-linus-4.9-rc1
> 
> The full pull request output is below.
> 
> Given the complexity and size of this change I am expecting - like
> the addition of reverse mapping last cycle - that there will be some
> follow-up bug fixes and cleanups around the -rc3 stage for issues
> that I'm sure will show up once the code hits a wider userbase.
> 
> What it is:
> 
> At the most basic level we are simply adding shared data extents to
> XFS - i.e. a single extent on disk can now have multiple owners. To
> do this we have to add new on-disk features to both track the shared
> extents and the number of times they've been shared. This is done by
> the new "refcount" btree that sits in every allocation group. When
> we share or unshare an extent, this tree gets updated.
> 
> Along with this new tree, the reverse mapping tree needs to be
> updated to track each owner or a shared extent. This also needs to
> be updated ever share/unshare operation. These interactions at
> extent allocation and freeing time have complex ordering and
> recovery constraints, so there's a significant amount of new
> intent-based transaction code to ensure that operations are
> performed atomically from both the runtime and integrity/crash
> recovery perspectives.
> 
> We also need to break sharing when writes hit a shared extent - this
> is where the new copy-on-write implementation comes in. We allocate
> new storage and copy the original data along with the overwrite data
> into the new location.  We only do this for data as we don't share
> metadata at all - each inode has it's own metadata that tracks the
> shared data extents, the extents undergoing CoW and it's own private
> extents.
> 
> Of course, being XFS, nothing is simple - we use delayed allocation
> for CoW similar to how we use it for normal writes. ENOSPC is a
> significant issue here - we build on the reservation code added
> in 4.8-rc1 with the reverse mapping feature to ensure we don't get
> spurious ENOSPC issues part way through a CoW operation. These
> mechanisms also help minimise fragmentation due to repeated CoW
> operations.  To further reduce fragmentation overhead, we've also
> introduced a CoW extent size hint, which indicates how large a
> region we should allocate when we execute a CoW operation.
> 
> With all this functionality in place, we can hook up
> .copy_file_range, .clone_file_range and .dedupe_file_range and we
> gain all the capabilities of reflink and other vfs provided
> functionality that enable manipulation to shared extents. We also
> added a fallocate mode that explicitly unshares a range of a file,
> which we implemented as an explicit CoW of all the shared extents in
> a file.
> 
> As such, it's a huge chunk of new functionality with new on-disk
> format features and internal infrastructure. It warns at mount time
> as an experimental feature and that it may eat data (as we do with
> all new on-disk features until they stabilise).  We have not
> released userspace suport for it yet - userspace support currently
> requires download from Darrick's xfsprogs repo and build from
> source, so the access to this feature is really developer/tester
> only at this point. Initial userspace support will be released at
> the same time the kernel with this code in it is released.

Userland support is in this branch:
https://github.com/djwong/xfsprogs/tree/for-dave-for-4.9-15

There will undoubtedly be more of these since Dave will libxfs-apply
the kernel patches into for-next after the merge window closes, after
which I'll rebase the tool patches against that.

> The new code causes 5-6 new failures with xfstests - these aren't
> serious functional failures but things the output of tests changing
> slightly due to perturbations in layouts, space usage, etc.  OTOH,
> we've added 150+ new tests to xfstests that specifically exercise
> this new functionality so it's got far better test coverage than any
> functionality we've previously added to XFS.

https://github.com/djwong/xfstests/tree/djwong-devel
have fixes to some of the tests tests, if you dare. :)

I'll resync with upstream the next time I see a xfstests.git update.
(Merge window is open, so I don't anticipate that until next week.)

> Darrick has done a pretty amazing job getting us to this stage, and
> special mention also needs to go to Christoph (review, testing,
> improvements and bug fixes) and Brian (caught several intricate
> bugs during review) for the effort they've also put in.

Yes, my hearty thanks to Dave, Christoph, and Brian for their support!

--D

> 
> Thanks,
> 
> -Dave.
> 
> ----------
> The following changes since commit 155cd433b516506df065866f3d974661f6473572:
> 
>   Merge branch 'xfs-4.9-log-recovery-fixes' into for-next (2016-10-03 09:56:28 +1100)
> 
> are available in the git repository at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs.git tags/xfs-reflink-for-linus-4.9-rc1
> 
> for you to fetch changes up to feac470e3642e8956ac9b7f14224e6b301b9219d:
> 
>   xfs: convert COW blocks to real blocks before unwritten extent conversion (2016-10-11 09:03:19 +1100)
> 
> ----------------------------------------------------------------
> xfs: reflink update for 4.9-rc1
> 
> < XFS has gained super CoW powers! >
>  ----------------------------------
>         \   ^__^
>          \  (oo)\_______
>             (__)\       )\/\
>                 ||----w |
>                 ||     ||
> 
> Included in this update:
> - unshare range (FALLOC_FL_UNSHARE) support for fallocate
> - copy-on-write extent size hints (FS_XFLAG_COWEXTSIZE) for fsxattr interface
> - shared extent support for XFS
> - copy-on-write support for shared extents
> - copy_file_range support
> - clone_file_range support (implements reflink)
> - dedupe_file_range support
> - defrag support for reverse mapping enabled filesystems
> 
> ----------------------------------------------------------------
> Christoph Hellwig (1):
>       xfs: convert COW blocks to real blocks before unwritten extent conversion
> 
> Darrick J. Wong (70):
>       vfs: support FS_XFLAG_COWEXTSIZE and get/set of CoW extent size hint
>       vfs: add a FALLOC_FL_UNSHARE mode to fallocate to unshare a range of blocks
>       xfs: return an error when an inline directory is too small
>       xfs: define tracepoints for refcount btree activities
>       xfs: introduce refcount btree definitions
>       xfs: refcount btree add more reserved blocks
>       xfs: define the on-disk refcount btree format
>       xfs: add refcount btree support to growfs
>       xfs: account for the refcount btree in the alloc/free log reservation
>       xfs: add refcount btree operations
>       xfs: create refcount update intent log items
>       xfs: log refcount intent items
>       xfs: adjust refcount of an extent of blocks in refcount btree
>       xfs: connect refcount adjust functions to upper layers
>       xfs: adjust refcount when unmapping file blocks
>       xfs: add refcount btree block detection to log recovery
>       xfs: reserve AG space for the refcount btree root
>       xfs: introduce reflink utility functions
>       xfs: create bmbt update intent log items
>       xfs: log bmap intent items
>       xfs: map an inode's offset to an exact physical block
>       xfs: pass bmapi flags through to bmap_del_extent
>       xfs: implement deferred bmbt map/unmap operations
>       xfs: when replaying bmap operations, don't let unlinked inodes get reaped
>       xfs: return work remaining at the end of a bunmapi operation
>       xfs: define tracepoints for reflink activities
>       xfs: add reflink feature flag to geometry
>       xfs: don't allow reflinked dir/dev/fifo/socket/pipe files
>       xfs: introduce the CoW fork
>       xfs: support bmapping delalloc extents in the CoW fork
>       xfs: create delalloc extents in CoW fork
>       xfs: support allocating delayed extents in CoW fork
>       xfs: allocate delayed extents in CoW fork
>       xfs: support removing extents from CoW fork
>       xfs: move mappings from cow fork to data fork after copy-write
>       xfs: report shared extent mappings to userspace correctly
>       xfs: implement CoW for directio writes
>       xfs: cancel CoW reservations and clear inode reflink flag when freeing blocks
>       xfs: cancel pending CoW reservations when destroying inodes
>       xfs: store in-progress CoW allocations in the refcount btree
>       xfs: reflink extents from one file to another
>       xfs: add clone file and clone range vfs functions
>       xfs: add dedupe range vfs function
>       xfs: teach get_bmapx about shared extents and the CoW fork
>       xfs: swap inode reflink flags when swapping inode extents
>       xfs: unshare a range of blocks via fallocate
>       xfs: create a separate cow extent size hint for the allocator
>       xfs: preallocate blocks for worst-case btree expansion
>       xfs: don't allow reflink when the AG is low on space
>       xfs: try other AGs to allocate a BMBT block
>       xfs: garbage collect old cowextsz reservations
>       xfs: increase log reservations for reflink
>       xfs: add shared rmap map/unmap/convert log item types
>       xfs: use interval query for rmap alloc operations on shared files
>       xfs: convert unwritten status of reverse mappings for shared files
>       xfs: set a default CoW extent size of 32 blocks
>       xfs: check for invalid inode reflink flags
>       xfs: don't mix reflink and DAX mode for now
>       xfs: simulate per-AG reservations being critically low
>       xfs: recognize the reflink feature bit
>       xfs: various swapext cleanups
>       xfs: refactor swapext code
>       xfs: implement swapext for rmap filesystems
>       xfs: check inode reflink flag before calling reflink functions
>       xfs: reduce stack usage of _reflink_clear_inode_flag
>       xfs: remove isize check from unshare operation
>       xfs: fix label inaccuracies
>       xfs: fix error initialization
>       xfs: clear reflink flag if setting realtime flag
>       xfs: rework refcount cow recovery error handling
> 
>  fs/open.c                          |    5 +
>  fs/xfs/Makefile                    |    7 +
>  fs/xfs/libxfs/xfs_ag_resv.c        |   15 +-
>  fs/xfs/libxfs/xfs_alloc.c          |   23 +
>  fs/xfs/libxfs/xfs_bmap.c           |  575 +++++++++++-
>  fs/xfs/libxfs/xfs_bmap.h           |   67 +-
>  fs/xfs/libxfs/xfs_bmap_btree.c     |   18 +
>  fs/xfs/libxfs/xfs_btree.c          |    8 +-
>  fs/xfs/libxfs/xfs_btree.h          |   16 +
>  fs/xfs/libxfs/xfs_defer.h          |    2 +
>  fs/xfs/libxfs/xfs_format.h         |   97 +-
>  fs/xfs/libxfs/xfs_fs.h             |   10 +-
>  fs/xfs/libxfs/xfs_inode_buf.c      |   24 +-
>  fs/xfs/libxfs/xfs_inode_buf.h      |    1 +
>  fs/xfs/libxfs/xfs_inode_fork.c     |   70 +-
>  fs/xfs/libxfs/xfs_inode_fork.h     |   28 +-
>  fs/xfs/libxfs/xfs_log_format.h     |  118 ++-
>  fs/xfs/libxfs/xfs_refcount.c       | 1698 ++++++++++++++++++++++++++++++++++++
>  fs/xfs/libxfs/xfs_refcount.h       |   70 ++
>  fs/xfs/libxfs/xfs_refcount_btree.c |  451 ++++++++++
>  fs/xfs/libxfs/xfs_refcount_btree.h |   74 ++
>  fs/xfs/libxfs/xfs_rmap.c           | 1120 +++++++++++++++++++++---
>  fs/xfs/libxfs/xfs_rmap.h           |    7 +
>  fs/xfs/libxfs/xfs_rmap_btree.c     |   82 +-
>  fs/xfs/libxfs/xfs_rmap_btree.h     |    7 +
>  fs/xfs/libxfs/xfs_sb.c             |    9 +
>  fs/xfs/libxfs/xfs_shared.h         |    2 +
>  fs/xfs/libxfs/xfs_trans_resv.c     |   23 +-
>  fs/xfs/libxfs/xfs_trans_resv.h     |    3 +
>  fs/xfs/libxfs/xfs_trans_space.h    |    9 +
>  fs/xfs/libxfs/xfs_types.h          |    3 +-
>  fs/xfs/xfs_aops.c                  |  222 ++++-
>  fs/xfs/xfs_aops.h                  |    4 +-
>  fs/xfs/xfs_bmap_item.c             |  508 +++++++++++
>  fs/xfs/xfs_bmap_item.h             |   98 +++
>  fs/xfs/xfs_bmap_util.c             |  589 ++++++++++---
>  fs/xfs/xfs_dir2_readdir.c          |    3 +-
>  fs/xfs/xfs_error.h                 |   10 +-
>  fs/xfs/xfs_file.c                  |  221 ++++-
>  fs/xfs/xfs_fsops.c                 |  107 ++-
>  fs/xfs/xfs_fsops.h                 |    3 +
>  fs/xfs/xfs_globals.c               |    5 +-
>  fs/xfs/xfs_icache.c                |  243 +++++-
>  fs/xfs/xfs_icache.h                |    7 +
>  fs/xfs/xfs_inode.c                 |   51 ++
>  fs/xfs/xfs_inode.h                 |   19 +
>  fs/xfs/xfs_inode_item.c            |    2 +-
>  fs/xfs/xfs_ioctl.c                 |   75 +-
>  fs/xfs/xfs_iomap.c                 |   35 +-
>  fs/xfs/xfs_iomap.h                 |    3 +-
>  fs/xfs/xfs_iops.c                  |    1 +
>  fs/xfs/xfs_itable.c                |    8 +-
>  fs/xfs/xfs_linux.h                 |    1 +
>  fs/xfs/xfs_log_recover.c           |  357 ++++++++
>  fs/xfs/xfs_mount.c                 |   32 +
>  fs/xfs/xfs_mount.h                 |    8 +
>  fs/xfs/xfs_ondisk.h                |    3 +
>  fs/xfs/xfs_pnfs.c                  |    7 +
>  fs/xfs/xfs_refcount_item.c         |  539 ++++++++++++
>  fs/xfs/xfs_refcount_item.h         |  101 +++
>  fs/xfs/xfs_reflink.c               | 1688 +++++++++++++++++++++++++++++++++++
>  fs/xfs/xfs_reflink.h               |   58 ++
>  fs/xfs/xfs_rmap_item.c             |   12 +
>  fs/xfs/xfs_stats.c                 |    1 +
>  fs/xfs/xfs_stats.h                 |   18 +-
>  fs/xfs/xfs_super.c                 |   87 ++
>  fs/xfs/xfs_sysctl.c                |    9 +
>  fs/xfs/xfs_sysctl.h                |    1 +
>  fs/xfs/xfs_trace.h                 |  742 +++++++++++++++-
>  fs/xfs/xfs_trans.h                 |   29 +
>  fs/xfs/xfs_trans_bmap.c            |  249 ++++++
>  fs/xfs/xfs_trans_refcount.c        |  264 ++++++
>  fs/xfs/xfs_trans_rmap.c            |    9 +
>  include/linux/falloc.h             |    3 +-
>  include/uapi/linux/falloc.h        |   18 +
>  include/uapi/linux/fs.h            |    4 +-
>  76 files changed, 10683 insertions(+), 413 deletions(-)
>  create mode 100644 fs/xfs/libxfs/xfs_refcount.c
>  create mode 100644 fs/xfs/libxfs/xfs_refcount.h
>  create mode 100644 fs/xfs/libxfs/xfs_refcount_btree.c
>  create mode 100644 fs/xfs/libxfs/xfs_refcount_btree.h
>  create mode 100644 fs/xfs/xfs_bmap_item.c
>  create mode 100644 fs/xfs/xfs_bmap_item.h
>  create mode 100644 fs/xfs/xfs_refcount_item.c
>  create mode 100644 fs/xfs/xfs_refcount_item.h
>  create mode 100644 fs/xfs/xfs_reflink.c
>  create mode 100644 fs/xfs/xfs_reflink.h
>  create mode 100644 fs/xfs/xfs_trans_bmap.c
>  create mode 100644 fs/xfs/xfs_trans_refcount.c
> -- 
> Dave Chinner
> david@...morbit.com
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists