linux-toradex.git/fs/xfs/xfs_buf.c, branch v6.9

fs,block: yield devices early

2024-03-27T12:17:15+00:00

Currently a device is only really released once the umount returns to
userspace due to how file closing works. That ultimately could cause
an old umount assumption to be violated that concurrent umount and mount
don't fail. So an exclusively held device with a temporary holder should
be yielded before the filesystem is gone. Add a helper that allows
callers to do that. This also allows us to remove the two holder ops
that Linus wasn't excited about.

Link: https://lore.kernel.org/r/20240326-vfs-bdev-end_holder-v1-1-20af85202918@kernel.org
Fixes: f3a608827d1f ("bdev: open block device as files") # mainline only
Reviewed-by: Christoph Hellwig 
Reviewed-by: Jan Kara 
Suggested-by: Linus Torvalds 
Signed-off-by: Christian Brauner

Merge tag 'xfs-6.9-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

2024-03-13T20:52:24+00:00

Pull xfs updates from Chandan Babu:

 - Online repair updates:
    - More ondisk structures being repaired:
        - Inode's mode field by trying to obtain file type value from
          the a directory entry
        - Quota counters
        - Link counts of inodes
        - FS summary counters
        - Support for in-memory btrees has been added to support repair
          of rmap btrees
    - Misc changes:
        - Report corruption of metadata to the health tracking subsystem
        - Enable indirect health reporting when resources are scarce
        - Reduce memory usage while repairing refcount btree
        - Extend "Bmap update" intent item to support atomic extent
          swapping on the realtime device
        - Extend "Bmap update" intent item to support extended attribute
          fork and unwritten extents
    - Code cleanups:
        - Bmap log intent
        - Btree block pointer checking
        - Btree readahead
        - Buffer target
        - Symbolic link code

 - Remove mrlock wrapper around the rwsem

 - Convert all the GFP_NOFS flag usages to use the scoped
   memalloc_nofs_save() API instead of direct calls with the GFP_NOFS

 - Refactor and simplify xfile abstraction. Lower level APIs in shmem.c
   are required to be exported in order to achieve this

 - Skip checking alignment constraints for inode chunk allocations when
   block size is larger than inode chunk size

 - Do not submit delwri buffers collected during log recovery when an
   error has been encountered

 - Fix SEEK_HOLE/DATA for file regions which have active COW extents

 - Fix lock order inversion when executing error handling path during
   shrinking a filesystem

 - Remove duplicate ifdefs

* tag 'xfs-6.9-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (183 commits)
  xfs: shrink failure needs to hold AGI buffer
  mm/shmem.c: Use new form of *@param in kernel-doc
  kernel-doc: Add unary operator * to $type_param_ref
  xfs: use kvfree() in xlog_cil_free_logvec()
  xfs: xfs_btree_bload_prep_block() should use __GFP_NOFAIL
  xfs: fix scrub stats file permissions
  xfs: fix log recovery erroring out on refcount recovery failure
  xfs: move symlink target write function to libxfs
  xfs: move remote symlink target read function to libxfs
  xfs: move xfs_symlink_remote.c declarations to xfs_symlink_remote.h
  xfs: xfs_bmap_finish_one should map unwritten extents properly
  xfs: support deferred bmap updates on the attr fork
  xfs: support recovering bmap intent items targetting realtime extents
  xfs: add a realtime flag to the bmap update log redo items
  xfs: add a xattr_entry helper
  xfs: fix xfs_bunmapi to allow unmapping of partial rt extents
  xfs: move xfs_bmap_defer_add to xfs_bmap_item.c
  xfs: reuse xfs_bmap_update_cancel_item
  xfs: add a bi_entry helper
  xfs: remove xfs_trans_set_bmap_flags
  ...

xfs: port block device access to files

2024-02-25T11:05:23+00:00

Link: https://lore.kernel.org/r/20240123-vfs-bdev-file-v2-7-adbd023e19cc@kernel.org
Reviewed-by: Christoph Hellwig 
Reviewed-by: Jan Kara 
Signed-off-by: Christian Brauner

xfs: support in-memory buffer cache targets

2024-02-22T20:43:21+00:00

Allow the buffer cache to target in-memory files by making it possible
to have a buftarg that maps pages from private shmem files.  As the
prevous patch alludes, the in-memory buftarg contains its own cache,
points to a shmem file, and does not point to a block_device.

The next few patches will make it possible to construct an xfs_btree in
pageable memory by using this buftarg.

Signed-off-by: Darrick J. Wong 
Reviewed-by: Christoph Hellwig

xfs: teach buftargs to maintain their own buffer hashtable

2024-02-22T20:42:58+00:00

Currently, cached buffers are indexed by per-AG hashtables.  This works
great for the data device, but won't work for in-memory btrees.  To
handle that use case, buftargs will need to be able to index buffers
independently of other data structures.

We accomplish this by hoisting the rhashtable and its lock into a
separate xfs_buf_cache structure, make the buftarg point to the
_buf_cache structure, and rework various functions to use it.  This
will enable the in-memory buftarg to come up with its own _buf_cache.

Signed-off-by: Darrick J. Wong 
Reviewed-by: Christoph Hellwig

xfs: move setting bt_logical_sectorsize out of xfs_setsize_buftarg

2024-02-22T20:42:45+00:00

bt_logical_sectorsize and the associated mask is set based on the
constant logical block size in the block_device structure and thus
doesn't need to be updated in xfs_setsize_buftarg.  Move it into
xfs_alloc_buftarg so that it is only done once per buftarg.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Darrick J. Wong

xfs: remove xfs_setsize_buftarg_early

2024-02-22T20:42:45+00:00

Open code the logic in the only caller, and improve the comment
explaining what is being done here.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Darrick J. Wong

xfs: remove the xfs_buftarg_t typedef

2024-02-22T20:42:44+00:00

Switch the few remaining holdouts to the struct version.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Darrick J. Wong

xfs: split xfs_buf_rele for cached vs uncached buffers

2024-02-22T20:41:02+00:00

xfs_buf_rele is a bit confusing because it mixes up handling of normal
cached and the special uncached buffers without much explanation.
Split the handling into two different helpers, and use a clearly named
helper that checks the hash key to distinguish the two cases instead
of checking the pag pointer.

Signed-off-by: Christoph Hellwig 
Reviewed-by: Darrick J. Wong 
Signed-off-by: Darrick J. Wong

xfs: clean up remaining GFP_NOFS users

2024-02-13T12:37:35+00:00

These few remaining GFP_NOFS callers do not need to use GFP_NOFS at
all. They are only called from a non-transactional context or cannot
be accessed from memory reclaim due to other constraints. Hence they
can just use GFP_KERNEL.

Signed-off-by: Dave Chinner 
Reviewed-by: "Darrick J. Wong" 
Signed-off-by: Chandan Babu R