Pull bfs2 fixes from Andreas Gruenbacher:
"Fix two bugs leading to leaked buffer head references:
- gfs2: Put bitmap buffers in put_super
- gfs2: Fix iomap buffer head reference counting bug
And one bug leading to significant slow-downs when deleting large
files:
- gfs2: Fix metadata read-ahead during truncate (2)"
* tag 'gfs2-4.20.fixes3' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
gfs2: Fix iomap buffer head reference counting bug
gfs2: Fix metadata read-ahead during truncate (2)
gfs2: Put bitmap buffers in put_super
GFS2 passes the inode buffer head (dibh) from gfs2_iomap_begin to
gfs2_iomap_end in iomap->private. It sets that private pointer in
gfs2_iomap_get. Users of gfs2_iomap_get other than gfs2_iomap_begin
would have to release iomap->private, but this isn't done correctly,
leading to a leak of buffer head references.
To fix this, move the code for setting iomap->private from
gfs2_iomap_get to gfs2_iomap_begin.
Fixes: 64bc06bb32 ("gfs2: iomap buffered write support")
Cc: stable@vger.kernel.org # v4.19+
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The previous attempt to fix for metadata read-ahead during truncate was
incorrect: for files with a height > 2 (1006989312 bytes with a block
size of 4096 bytes), read-ahead requests were not being issued for some
of the indirect blocks discovered while walking the metadata tree,
leading to significant slow-downs when deleting large files. Fix that.
In addition, only issue read-ahead requests in the first pass through
the meta-data tree, while deallocating data blocks.
Fixes: c3ce5aa9b0 ("gfs2: Fix metadata read-ahead during truncate")
Cc: stable@vger.kernel.org # v4.16+
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
1. Andreas Gruenbacher contributed several patches to clean up the gfs2
block allocator to prepare for future performance enhancements.
2. Andy Price contributed a patch to fix a use-after-free problem.
3. I contributed some patches that fix gfs2's broken rgrplvb mount option.
4. I contributed some cleanup patches and error message improvements.
5. Steve Whitehouse and Abhi Das sent a patch to enable getlabel support.
6. Tim Smith contributed a patch to flush the glock delete workqueue at exit.
-----BEGIN PGP SIGNATURE-----
iQEcBAABAgAGBQJbz3QlAAoJENeLYdPf93o7+VUH/0XUfbyNQsU6fyLP8NrZq05z
qNsVN3Hm+tPc0/V0C75lSp9ej7B8ogMl0RPysniziRTEWIDK6oGB/JUGvIH0O/z1
vZ/sofZEDXthV3YjiI8RXLcaJLsOavSXnGwHbNKohM2PdRObVkZbaUL+xWlL9X3q
yHgP5AHCIrpVzz5l4sLO6N0Npnl0aNRTBxPIyDTaBBmitXkvtqkCbkw185jzkDDs
fMvZ6I+3UbUxp99InFTHeUXvTr1EbvfPrhZzmppuV1N4LLSa1eRaWmsKTEDPdnsy
uhsh9ittv8EJXN2dpmZIOdGmDEK07kFoZrsbM5F78sOH/LbUyJ5YfBN02lDaaAI=
=b7zk
-----END PGP SIGNATURE-----
Merge tag 'gfs2-4.20.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull gfs2 updates from Bob Peterson:
"We've got 18 patches for this merge window, none of which are very
major:
- clean up the gfs2 block allocator to prepare for future performance
enhancements (Andreas Gruenbacher)
- fix a use-after-free problem (Andy Price)
- patches that fix gfs2's broken rgrplvb mount option (me)
- cleanup patches and error message improvements (me)
- enable getlabel support (Steve Whitehouse and Abhi Das)
- flush the glock delete workqueue at exit (Tim Smith)"
* tag 'gfs2-4.20.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
gfs2: Fix minor typo: couln't versus couldn't.
gfs2: write revokes should traverse sd_ail1_list in reverse
gfs2: Pass resource group to rgblk_free
gfs2: Remove unnecessary gfs2_rlist_alloc parameter
gfs2: Fix marking bitmaps non-full
gfs2: Fix some minor typos
gfs2: Rename bitmap.bi_{len => bytes}
gfs2: Remove unused RGRP_RSRV_MINBYTES definition
gfs2: Move rs_{sizehint, rgd_gh} fields into the inode
gfs2: Clean up out-of-bounds check in gfs2_rbm_from_block
gfs2: Always check the result of gfs2_rbm_from_block
gfs2: getlabel support
GFS2: Flush the GFS2 delete workqueue before stopping the kernel threads
gfs2: Don't leave s_fs_info pointing to freed memory in init_sbd
gfs2: Use fs_* functions instead of pr_* function where we can
gfs2: slow the deluge of io error messages
gfs2: Don't set GFS2_RDF_UPTODATE when the lvb is updated
gfs2: improve debug information when lvb mismatches are found
It turns out that the fix in commit 6636c3cc56 is bad; the assertion
that the iomap code no longer creates buffer heads is incorrect for
filesystems that set the IOMAP_F_BUFFER_HEAD flag.
Instead, what's happening is that gfs2_iomap_begin_write treats all
files that have the jdata flag set as journaled files, which is
incorrect as long as those files are inline ("stuffed"). We're handling
stuffed files directly via the page cache, which is why we ended up with
pages without buffer heads in gfs2_page_add_databufs.
Fix this by handling stuffed journaled files correctly in
gfs2_iomap_begin_write.
This reverts commit 6636c3cc5690c11631e6366cf9a28fb99c8b25bb.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Function rgblk_free can only deal with one resource group at a time, so
pass that resource group is as a parameter. Several of the callers
already have the resource group at hand, so we only need additional
lookup code in a few places.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Reviewed-by: Steven Whitehouse <swhiteho@redhat.com>
Commit 64bc06bb32 broke buffered writes to journaled files (chattr
+j): we'll try to journal the buffer heads of the page being written to
in gfs2_iomap_journaled_page_done. However, the iomap code no longer
creates buffer heads, so we'll BUG() in gfs2_page_add_databufs. Fix
that by creating buffer heads ourself when needed.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
To speed up the common case of appending to a file,
gfs2_write_alloc_required presumes that writing beyond the end of a file
will always require additional blocks to be allocated. This assumption
is incorrect for preallocates files, but there are no negative
consequences as long as *some* space is still left on the filesystem.
One special file that always has some space preallocated beyond the end
of the file is the rindex: when growing a filesystem, gfs2_grow adds one
or more new resource groups and appends records describing those
resource groups to the rindex; the preallocated space ensures that this
is always possible.
However, when a filesystem is completely full, gfs2_write_alloc_required
will indicate that an additional allocation is required, and appending
the next record to the rindex will fail even though space for that
record has already been preallocated. To fix that, skip the incorrect
optimization in gfs2_write_alloc_required, but for the rindex only.
Other writes to preallocated space beyond the end of the file are still
allowed to fail on completely full filesystems.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
Merge xfs branch 'iomap-4.19-merge' into linux-gfs2/for-next. This
brings in readpage and direct I/O support for inline data.
The IOMAP_F_BUFFER_HEAD flag introduced in commit "iomap: add initial
support for writes without buffer heads" needs to be set for gfs2 as
well, so do that in the merge.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Pull in the gfs2 iomap-write changes: Tweak the existing code to
properly support iomap write and eliminate an unnecessary special case
in gfs2_block_map. Implement iomap write support for buffered and
direct I/O. Simplify some of the existing code and eliminate code that
is no longer used:
gfs2: Remove gfs2_write_{begin,end}
gfs2: iomap direct I/O support
gfs2: gfs2_extent_length cleanup
gfs2: iomap buffered write support
gfs2: Further iomap cleanups
This is based on the following changes on the xfs 'iomap-4.19-merge'
branch:
iomap: add private pointer to struct iomap
iomap: add a page_done callback
iomap: generic inline data handling
iomap: complete partial direct I/O writes synchronously
iomap: mark newly allocated buffer heads as new
fs: factor out a __generic_write_end helper
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
The page unmapping previously done in gfs2_direct_IO is now done
generically in iomap_dio_rw.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
Now that gfs2_extent_length is no longer used for determining the size
of a hole and always with an upper size limit, the function can be
simplified.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
With the traditional page-based writes, blocks are allocated separately
for each page written to. With iomap writes, we can allocate a lot more
blocks at once, with a fraction of the allocation overhead for each
page.
Split calculating the number of blocks that can be allocated at a given
position (gfs2_alloc_size) off from gfs2_iomap_alloc: that size
determines the number of blocks to allocate and reserve in the journal.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
In gfs2_iomap_alloc, set the type of newly allocated extents to
IOMAP_MAPPED so that iomap_to_bh will set the bh states correctly:
otherwise, the bhs would not be marked as mapped, confusing
__mpage_writepage. This means that we need to check for the IOMAP_F_NEW
flag in fallocate_chunk now.
Further clean up gfs2_iomap_get and implement gfs2_stuffed_iomap here
directly. For reads beyond the end of the file, return holes instead of
failing with -ENOENT so that we can get rid of that special case in
gfs2_block_map.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
Rename end_off to end_len to make the code less confusing.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
- Strengthen inode number and structure validation when allocating inodes.
- Reduce pointless buffer allocations during cache miss
- Use FUA for pure data O_DSYNC directio writes
- Various iomap refactorings
- Strengthen quota metadata verification to avoid unfixable broken quota
- Make AGFL block freeing a deferred operation to avoid blowing out
transaction reservations when running complex operations
- Get rid of the log item descriptors to reduce log overhead
- Fix various reflink bugs where inodes were double-joined to
transactions
- Don't issue discards when trimming unwritten extents
- Refactor incore dquot initialization and retrieval interfaces
- Fix some locking problmes in the quota scrub code
- Strengthen btree structure checks in scrub code
- Rewrite swapfile activation to use iomap and support unwritten extents
- Make scrub exit to userspace sooner when corruptions or
cross-referencing problems are found
- Make scrub invoke the data fork scrubber directly on metadata inodes
- Don't do background reclamation of post-eof and cow blocks when the fs
is suspended
- Fix secondary superblock buffer lifespan hinting
- Refactor growfs to use table-dispatched functions instead of long
stringy functions
- Move growfs code to libxfs
- Implement online fs label getting and setting
- Introduce online filesystem repair (in a very limited capacity)
- Fix unit conversion problems in the realtime freemap iteration
functions
- Various refactorings and cleanups in preparation to remove buffer
heads in a future release
- Reimplement the old bmap call with iomap
- Remove direct buffer head accesses from seek hole/data
- Various bug fixes
-----BEGIN PGP SIGNATURE-----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=11pr
-----END PGP SIGNATURE-----
Merge tag 'xfs-4.18-merge-3' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull xfs updates from Darrick Wong:
"New features this cycle include the ability to relabel mounted
filesystems, support for fallocated swapfiles, and using FUA for pure
data O_DSYNC directio writes. With this cycle we begin to integrate
online filesystem repair and refactor the growfs code in preparation
for eventual subvolume support, though the road ahead for both
features is quite long.
There are also numerous refactorings of the iomap code to remove
unnecessary log overhead, to disentangle some of the quota code, and
to prepare for buffer head removal in a future upstream kernel.
Metadata validation continues to improve, both in the hot path
veifiers and the online filesystem check code. I anticipate sending a
second pull request in a few days with more metadata validation
improvements.
This series has been run through a full xfstests run over the weekend
and through a quick xfstests run against this morning's master, with
no major failures reported.
Summary:
- Strengthen inode number and structure validation when allocating
inodes.
- Reduce pointless buffer allocations during cache miss
- Use FUA for pure data O_DSYNC directio writes
- Various iomap refactorings
- Strengthen quota metadata verification to avoid unfixable broken
quota
- Make AGFL block freeing a deferred operation to avoid blowing out
transaction reservations when running complex operations
- Get rid of the log item descriptors to reduce log overhead
- Fix various reflink bugs where inodes were double-joined to
transactions
- Don't issue discards when trimming unwritten extents
- Refactor incore dquot initialization and retrieval interfaces
- Fix some locking problmes in the quota scrub code
- Strengthen btree structure checks in scrub code
- Rewrite swapfile activation to use iomap and support unwritten
extents
- Make scrub exit to userspace sooner when corruptions or
cross-referencing problems are found
- Make scrub invoke the data fork scrubber directly on metadata
inodes
- Don't do background reclamation of post-eof and cow blocks when the
fs is suspended
- Fix secondary superblock buffer lifespan hinting
- Refactor growfs to use table-dispatched functions instead of long
stringy functions
- Move growfs code to libxfs
- Implement online fs label getting and setting
- Introduce online filesystem repair (in a very limited capacity)
- Fix unit conversion problems in the realtime freemap iteration
functions
- Various refactorings and cleanups in preparation to remove buffer
heads in a future release
- Reimplement the old bmap call with iomap
- Remove direct buffer head accesses from seek hole/data
- Various bug fixes"
* tag 'xfs-4.18-merge-3' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (121 commits)
fs: use ->is_partially_uptodate in page_cache_seek_hole_data
fs: remove the buffer_unwritten check in page_seek_hole_data
fs: move page_cache_seek_hole_data to iomap.c
xfs: use iomap_bmap
iomap: add an iomap-based bmap implementation
iomap: add a iomap_sector helper
iomap: use __bio_add_page in iomap_dio_zero
iomap: move IOMAP_F_BOUNDARY to gfs2
iomap: fix the comment describing IOMAP_NOWAIT
iomap: inline data should be an iomap type, not a flag
mm: split ->readpages calls to avoid non-contiguous pages lists
mm: return an unsigned int from __do_page_cache_readahead
mm: give the 'ret' variable a better name __do_page_cache_readahead
block: add a lower-level bio_add_page interface
xfs: fix error handling in xfs_refcount_insert()
xfs: fix xfs_rtalloc_rec units
xfs: strengthen rtalloc query range checks
xfs: xfs_rtbuf_get should check the bmapi_read results
xfs: xfs_rtword_t should be unsigned, not signed
dax: change bdev_dax_supported() to support boolean returns
...
Clean up gfs2_iomap_alloc and gfs2_iomap_get. Document how
gfs2_iomap_alloc works: it now needs to be called separately after
gfs2_iomap_get where necessary; this will be used later by iomap write.
Move gfs2_iomap_ops into bmap.c.
Introduce a new gfs2_iomap_get_alloc helper and use it in
fallocate_chunk: gfs2_iomap_begin will become unsuitable for fallocate
with proper iomap write support.
In gfs2_block_map and fallocate_chunk, zero-initialize struct iomap.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
In journaled data mode, we need to add each buffer head to the current
transaction. In ordered write mode, we only need to add the inode to
the ordered inode list. So far, both cases are handled in
gfs2_trans_add_data. This makes the code look misleading and is
inefficient for small block sizes as well. Handle both cases separately
instead.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Reimplement function hole_size based on a generic function for walking
the metadata tree and rename hole_size to gfs2_hole_size. While
previously, multiple invocations of hole_size were sometimes needed to
walk across the entire hole, the new implementation always returns the
entire hole at once (provided that the caller is interested in the total
size).
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Just define a range of fs specific flags and use that in gfs2 instead of
exposing this internal flag globally.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Inline data is fundamentally different from our normal mapped case in that
it doesn't even have a block address. So instead of having a flag for it
it should be an entirely separate iomap range type.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
GFS2 keeps two arrarys in the superblock that define the maximum size of
an inode depending on the inode's height: sdp->sd_heightsize defines the
heights in units of sb->s_blocksize; sdp->sd_jheightsize defines them in
units of sb->s_blocksize - sizeof(struct gfs2_meta_header). These
arrays are used to determine when additional layers of indirect blocks
are needed. The second array is used for directories which have an
additional gfs2_meta_header at the beginning of each block.
Distinguishing between these two cases makes no sense: the height
required for representing N blocks will come out the same no matter if
the calculation is done in gross (sb->s_blocksize) or net
(sb->s_blocksize - sizeof(struct gfs2_meta_header)) units.
Stuffed directories don't have an additional gfs2_meta_header, but the
stuffed case is handled separately for both files and directories,
anyway.
Remove the unncessary sdp->sd_jheightsize array.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
This patch simply fixes some comments and the gfs2-glocks.txt file:
Places where i_rwsem was called i_mutex, and adding i_rw_mutex.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Instead of zeroing out fallocated blocks in gfs2_iomap_alloc, zero them
out in fallocate_chunk, much higher up the call stack. This gets rid of
gfs2's abuse of the IOMAP_ZERO flag as well as the gfs2 specific zeronew
buffer flag. I can't think of a reason why zeroing out the blocks in
gfs2_iomap_alloc would have any benefits: there is no additional locking
at that level that would add protection to the newly allocated blocks.
While at it, change fallocate over from gs2_block_map to gfs2_iomap_begin.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Acked-by: Christoph Hellwig <hch@lst.de>
When punching a hole or truncating an inode down to a given size, also
check if the truncate point / start of the hole is within the range we
have metadata for. Otherwise, we can end up freeing blocks that
shouldn't be freed, corrupting the inode, or crashing the machine when
trying to punch a hole into the void.
When growing an inode via truncate, we set the new size but we don't
allocate additional levels of indirect blocks and grow the inode height.
When shrinking that inode again, the new size may still point beyond the
end of the inode's metadata.
Fixes xfstest generic/476.
Debugged-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
It turns out that commit 3229c18c0d6b2 'Fixes to "Implement iomap for
block_map"' introduced another bug in gfs2_iomap_begin that can cause
gfs2_block_map to set bh->b_size of an actual buffer to 0. This can
lead to arbitrary incorrect behavior including crashes or disk
corruption. Revert the incorrect part of that commit.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
It turns out that commit 3974320ca6 "Implement iomap for block_map"
introduced a few bugs that trigger occasional failures with xfstest
generic/476:
In gfs2_iomap_begin, we jump to do_alloc when we determine that we are
beyond the end of the allocated metadata (height > ip->i_height).
There, we can end up calling hole_size with a metapath that doesn't
match the current metadata tree, which doesn't make sense. After
untangling the code at do_alloc, fix this by checking if the block we
are looking for is within the range of allocated metadata.
In addition, add a BUG() in case gfs2_iomap_begin is accidentally called
for reading stuffed files: this is handled separately. Make sure we
don't truncate iomap->length for reads beyond the end of the file; in
that case, the entire range counts as a hole.
Finally, revert to taking a bitmap write lock when doing allocations.
It's unclear why that change didn't lead to any failures during testing.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Add a small inline function for computing the maximum size of a stuffed
inode instead of open coding that in several places throughout the code.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Implement the top-level bits of punching a hole into a file.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Add an upper bound to the range of blocks to deallocate blocks to
function trunc_dealloc so that this function can be used for truncating
a file as well as for punching a hole into a file.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Pull the code for computing the range of metapointers to iterate out of
gfs2_metapath_ra (for readahead), sweep_bh_for_rgrps (for deallocating
metapointers within a block), and trunc_dealloc (for walking the
metadata tree).
In sweep_bh_for_rgrps, move the code for looking up the resource group
descriptor of the current resource group out of the inner loop. The
metatype check moves to trunc_dealloc.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Turn gfs2_block_truncate_page into a function that zeroes a range within
a block rather than only the end of a block. This will be used for
cleaning the end of the first partial block and the start of the last
partial block when punching a hole in a file.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
In rare cases, the current non-recursive delete algorithm doesn't
deallocate empty intermediary indirect blocks. This should have very
little practical effect, but deallocating all blocks correctly should
still be preferable as it is cleaner and easier to validate.
The fix consists of using the first block to deallocate to compute the
start marker of the truncate point instead of the last block that needs
to be kept. With that change, computing which indirect blocks are still
needed becomes relatively easy.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
The metadata read-ahead algorithm broke when switching from recursive to
non-recursive delete: the current algorithm reads ahead blocks at height
N - 1 while deallocating the blocks at hight N. However, deallocating
the blocks at height N requires a complete walk of the metadata tree,
not only down to height N - 1. Consequently, all blocks below height
N - 1 will be accessed without read-ahead.
Fix this by issuing read-aheads as early as possible, after each
metapath lookup.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Split out the entire lookup loop from lookup_metapath and
fillup_metapath. Make both functions return the actual height in
mp->mp_aheight, and return 0 on success. Handle lookup errors properly
in trunc_dealloc.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
First, this function truncates the file in chunks. When the original
file size isn't block aligned, each chunk that is truncated will remain
be misaligned. This is inefficient.
Second, this function doesn't recognize where holes are, so it loops
through them. For each chunk of a hole, it creates a new transaction.
At least avoid creating another transactions whe the current one is
still empty. (An better fix would be to skip large holes, of course.)
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Document when to use gfs2_blk2rgrpd for "inexact" resource group
matching. Based on that, fix an incorrect use of gfs2_blk2rgrpd in
sweep_bh_for_rgrps.
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
This patch implements iomap for block mapping, and switches the
block_map function to use it under the covers.
The additional IOMAP_F_BOUNDARY iomap flag indicates when iomap has
reached a "metadata boundary" and fetching the next mapping is likely to
incur an additional I/O. This flag is used for setting the bh buffer
boundary flag.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
This patch eliminates height parameters from function gfs2_bmap_alloc.
Function find_metapath determines the metapath's "find height", also
known as the desired height. Function lookup_metapath determines the
metapath's "actual height", previously known as starting height or
sheight. Function gfs2_bmap_alloc now gets both height values from
the metapath. This simplification was done as a step toward switching
the block_map functions to using iomap. The bh_map responsibilities
are also removed from function gfs2_bmap_alloc for the same reason.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Add a comment about the logical block size for directories. Rename
"bsize" in gfs2_block_map to "factor". Fix a typo in the description of
metaptr1.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Before this patch if you truncated a file to a smaller size it
wasn't freeing all the blocks properly. There are two reasons.
First, the metapath comparison was not comparing previous heights.
I added a function, mp_eq_to_hgt, which checks the metapath at
all heights prior to the target height.
Second, in function find_nonnull_ptr, it needed to zero out all
pointers for heights following the target height. Translated into
decimal integer terms, this way a number like 299, when incremented,
becomes 300, not 399. The 2 gets incremented to 3, and the following
digits need to be reset.
These two things allow the truncate state machine to properly find
the blocks it needs to delete.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
When gfs2 does metadata I/O, only REQ_META is used as a metadata hint of
the bio. But flag REQ_META is just a hint for block trace, not for block
layer code to handle a bio as metadata request.
For some of metadata I/Os of gfs2, A REQ_PRIO flag on the metadata bio
would be very informative to block layer code. For example, if bcache is
used as a I/O cache for gfs2, it will be possible for bcache code to get
the hint and cache the pre-fetched metadata blocks on cache device. This
behavior may be helpful to improve metadata I/O performance if the
following requests hit the cache.
Here are the locations in gfs2 code where a REQ_PRIO flag should be added,
- All places where REQ_READAHEAD is used, gfs2 code uses this flag for
metadata read ahead.
- In gfs2_meta_rq() where the first metadata block is read in.
- In gfs2_write_buf_to_page(), read in quota metadata blocks to have them
up to date.
These metadata blocks are probably to be accessed again in future, adding
a REQ_PRIO flag may have bcache to keep such metadata in fast cache
device. For system without a cache layer, REQ_PRIO can still provide hint
to block layer to handle metadata requests more properly.
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
1. Andreas Gruenbacher has four patches related to cleaning up the GFS2
inode evict process. This is about half of his patches designed to
fix a long-standing GFS2 hang related to the inode shrinker.
(Shrinker calls gfs2 evict, evict calls DLM, DLM requires memory
and blocks on the shrinker.) These 4 patches have been well tested.
His second set of patches are still being tested, so I plan to hold
them until the next merge window, after we have more weeks of testing.
The first patch eliminates the flush_delayed_work, which can block.
2. Andreas's second patch protects setting of gl_object for rgrps with
a spin_lock to prevent proven races.
3. His third patch introduces a centralized mechanism for queueing glock
work with better reference counting, to prevent more races.
4. His fourth patch retains a reference to inode glocks when an error
occurs while creating an inode. This keeps the subsequent evict from
needing to reacquire the glock, which might call into DLM and block
in low memory conditions.
5. Arvind Yadav has a patch to add const to attribute_group structures.
6. I have a patch to detect directory entry inconsistencies and withdraw
the file system if any are found. Better that than silent corruption.
7. I have a patch to remove a vestigial variable from glock structures,
saving some slab space.
8. I have another patch to remove a vestigial variable from the GFS2
in-core superblock structure.
-----BEGIN PGP SIGNATURE-----
iQEcBAABAgAGBQJZXOIfAAoJENeLYdPf93o7RVcH/jLEK3hmZOd94pDTYg3Damuo
KI3xjyutDgQT83uwg8p5UBPwRYCDnyiOLwOWGBJJvjPEI1S4syrXq/FzOmxmX6cV
nE28ARL/OXCoFEXBMUVHvHL3nK+zEUr8rO6Xz51B1ifVq7GV8iVK+ZgxzRhx0PWP
f+0SVHiQtU0HKyxR5y9p43oygtHZaGbjy4WL0YbmFZM59y5q9A8rBHFACn2JyPBm
/zXN6gF/Orao+BDXLT6OM3vNXZcOQ7FUPWwctguHsAO/bLzWiISyfJxLWJsHvSdW
tzFTN1DByjXvqAhs4HTSuh9JfBDAyxcXkmczXJyATBkCTEJv42Iev+ILmre+wwQ=
=YTwn
-----END PGP SIGNATURE-----
Merge tag 'gfs2-4.13.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull GFS2 updates from Bob Peterson:
"We've got eight GFS2 patches for this merge window:
- Andreas Gruenbacher has four patches related to cleaning up the
GFS2 inode evict process. This is about half of his patches
designed to fix a long-standing GFS2 hang related to the inode
shrinker: Shrinker calls gfs2 evict, evict calls DLM, DLM requires
memory and blocks on the shrinker.
These four patches have been well tested. His second set of patches
are still being tested, so I plan to hold them until the next merge
window, after we have more weeks of testing. The first patch
eliminates the flush_delayed_work, which can block.
- Andreas's second patch protects setting of gl_object for rgrps with
a spin_lock to prevent proven races.
- His third patch introduces a centralized mechanism for queueing
glock work with better reference counting, to prevent more races.
-His fourth patch retains a reference to inode glocks when an error
occurs while creating an inode. This keeps the subsequent evict
from needing to reacquire the glock, which might call into DLM and
block in low memory conditions.
- Arvind Yadav has a patch to add const to attribute_group
structures.
- I have a patch to detect directory entry inconsistencies and
withdraw the file system if any are found. Better that than silent
corruption.
- I have a patch to remove a vestigial variable from glock
structures, saving some slab space.
- I have another patch to remove a vestigial variable from the GFS2
in-core superblock structure"
* tag 'gfs2-4.13.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
GFS2: constify attribute_group structures.
gfs2: gfs2_create_inode: Keep glock across iput
gfs2: Clean up glock work enqueuing
gfs2: Protect gl->gl_object by spin lock
gfs2: Get rid of flush_delayed_work in gfs2_evict_inode
GFS2: Eliminate vestigial sd_log_flush_wrapped
GFS2: Remove gl_list from glock structure
GFS2: Withdraw when directory entry inconsistencies are detected
Put all remaining accesses to gl->gl_object under the
gl->gl_lockref.lock spinlock to prevent races.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Implement truncate/delete as a non-recursive algorithm. The older
algorithm was implemented with recursion to strip off each layer
at a time (going by height, starting with the maximum height.
This version tries to do the same thing but without recursion,
and without needing to allocate new structures or lists in memory.
For example, say you want to truncate a very large file to 1 byte,
and its end-of-file metapath is: 0.505.463.428. The starting
metapath would be 0.0.0.0. Since it's a truncate to non-zero, it
needs to preserve that byte, and all metadata pointing to it.
So it would start at 0.0.0.0, look up all its metadata buffers,
then free all data blocks pointed to at the highest level.
After that buffer is "swept", it moves on to 0.0.0.1, then
0.0.0.2, etc., reading in buffers and sweeping them clean.
When it gets to the end of the 0.0.0 metadata buffer (for 4K
blocks the last valid one is 0.0.0.508), it backs up to the
previous height and starts working on 0.0.1.0, then 0.0.1.1,
and so forth. After it reaches the end and sweeps 0.0.1.508,
it continues with 0.0.2.0, and so on. When that height is
exhausted, and it reaches 0.0.508.508 it backs up another level,
to 0.1.0.0, then 0.1.0.1, through 0.1.0.508. So it has to keep
marching backwards and forwards through the metadata until it's
all swept clean. Once it has all the data blocks freed, it
lowers the strip height, and begins the process all over again,
but with one less height. This time it sweeps 0.0.0 through
0.505.463. When that's clean, it lowers the strip height again
and works to free 0.505. Eventually it strips the lowest height, 0.
For a delete or truncate to 0, all metadata for all heights of
0.0.0.0 would be freed. For a truncate to 1 byte, 0.0.0.0 would
be preserved.
This isn't much different from normal integer incrementing,
where an integer gets incremented from 0000 (0.0.0.0) to 3021
(3.0.2.1). So 0000 gets increments to 0001, 0002, up to 0009,
then on to 0010, 0011 up to 0099, then 0100 and so forth. It's
just that each "digit" goes from 0 to 508 (for a total of 509
pointers) rather than from 0 to 9.
Note that the dinode will only have 483 pointers due to the
dinode structure itself.
Also note: this is just an example. These numbers (509 and 483)
are based on a standard 4K block size. Smaller block sizes will
yield smaller numbers of indirect pointers accordingly.
The truncation process is accomplished with the help of two
major functions and a few helper functions.
Functions do_strip and recursive_scan are obsolete, so removed.
New function sweep_bh_for_rgrps cleans a buffer_head pointed to
by the given metapath and height. By cleaning, I mean it frees
all blocks starting at the offset passed in metapath. It starts
at the first block in the buffer pointed to by the metapath and
identifies its resource group (rgrp). From there it frees all
subsequent block pointers that lie within that rgrp. If it's
already inside a transaction, it stays within it as long as it
can. In other words, it doesn't close a transaction until it knows
it's freed what it can from the resource group. In this way,
multiple buffers may be cleaned in a single transaction, as long
as those blocks in the buffer all lie within the same rgrp.
If it's not in a transaction, it starts one. If the buffer_head
has references to blocks within multiple rgrps, it frees all the
blocks inside the first rgrp it finds, then closes the
transaction. Then it repeats the cycle: identifies the next
unfreed block, uses it to find its rgrp, then starts a new
transaction for that set. It repeats this process repeatedly
until the buffer_head contains no more references to any blocks
past the given metapath.
Function trunc_dealloc has been reworked into a finite state
automaton. It has basically 3 active states:
DEALLOC_MP_FULL, DEALLOC_MP_LOWER, and DEALLOC_FILL_MP:
The DEALLOC_MP_FULL state implies the metapath has a full set
of buffers out to the "shrink height", and therefore, it can
call function sweep_bh_for_rgrps to free the blocks within the
highest height of the metapath. If it's just swept the lowest
level (or an error has occurred) the state machine is ended.
Otherwise it proceeds to the DEALLOC_MP_LOWER state.
The DEALLOC_MP_LOWER state implies we are finished with a given
buffer_head, which may now be released, and therefore we are
then missing some buffer information from the metapath. So we
need to find more buffers to read in. In most cases, this is
just a matter of releasing the buffer_head and moving to the
next pointer from the previous height, so it may be read in and
swept as well. If it can't find another non-null pointer to
process, it checks whether it's reached the end of a height
and needs to lower the strip height, or whether it still needs
move forward through the previous height's metadata. In this
state, all zero-pointers are skipped. From this state, it can
only loop around (once more backing up another height) or,
once a valid metapath is found (one that has non-zero
pointers), proceed to state DEALLOC_FILL_MP.
The DEALLOC_FILL_MP state implies that we have a metapath
but not all its buffers are read in. So we must proceed to read
in buffer_heads until the metapath has a valid buffer for every
height. If the previous state backed us up 3 heights, we may
need to read in a buffer, increment the height, then repeat the
process until buffers have been read in for all required heights.
If it's successful reading a buffer, and it's at the highest
height we need, it proceeds back to the DEALLOC_MP_FULL state.
If it's unable to fill in a buffer, (encounters a hole, etc.)
it tries to find another non-zero block pointer. If they're all
zero, it lowers the height and returns to the DEALLOC_MP_LOWER
state. If it finds a good non-null pointer, it loops around and
reads it in, while keeping the metapath in lock-step with the
pointers it examines.
The state machine runs until the truncation request is
satisfied. Then any transactions are ended, the quota and
statfs data are updated, and the function is complete.
Helper function metaptr1 was introduced to be an easy way to
determine the start of a buffer_head's indirect pointers.
Helper function lookup_mp_height was introduced to find a
metapath index and read in the buffer that corresponds to it.
In this way, function lookup_metapath becomes a simple loop to
call it for every height.
Helper function fillup_metapath is similar to lookup_metapath
except it can do partial lookups. If the state machine
backed up multiple levels (like 2999 wrapping to 3000) it
needs to find out the next starting point and start issuing
metadata reads at that point.
Helper function hptrs is a shortcut to determine how many
pointers should be expected in a buffer. Height 0 is the dinode
which has fewer pointers than the others.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
