Recently, we found that the CPU spent a lot of time in
xfs_alloc_ag_vextent_size when the filesystem has millions of fragmented
spaces.
The reason is that we conducted much extra searching for extents that
could not yield a better result, and these searches would cost a lot of
time when there were millions of extents to search through. Even if we
get the same result length, we don't switch our choice to the new one,
so we can definitely terminate the search early.
Since the result length cannot exceed the found length, when the found
length equals the best result length we already have, we can conclude
the search.
We did a test in that filesystem:
[root@localhost ~]# xfs_db -c freesp /dev/vdb
from to extents blocks pct
1 1 215 215 0.01
2 3 994476 1988952 99.99
Before this patch:
0) | xfs_alloc_ag_vextent_size [xfs]() {
0) * 15597.94 us | }
After this patch:
0) | xfs_alloc_ag_vextent_size [xfs]() {
0) 19.176 us | }
Signed-off-by: Chi Zhiling <chizhiling@kylinos.cn>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Currently log recovery never updates the in-core perag values for the
last allocation group when they were grown by growfs. This leads to
btree record validation failures for the alloc, ialloc or finotbt
trees if a transaction references this new space.
Found by Brian's new growfs recovery stress test.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
__GFP_RETRY_MAYFAIL increases the likelyhood of allocations to fail,
which isn't really helpful during log recovery. Remove the flag and
stick to the default GFP_KERNEL policies.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
There is no good reason to have two different routines for freeing perag
structures for the unmount and error cases. Add two arguments to specify
the range of AGs to free to xfs_free_perag, and use that to replace
xfs_free_unused_perag_range.
The addition RCU grace period for the error case is harmless, and the
extra check for the AG to actually exist is not required now that the
callers pass the exact known allocated range.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Currently only the new agcount is passed to xfs_initialize_perag, which
requires lookups of existing AGs to skip them and complicates error
handling. Also pass the previous agcount so that the range that
xfs_initialize_perag operates on is exactly defined. That way the
extra lookups can be avoided, and error handling can clean up the
exact range from the old count to the last added perag structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Jeff Layton <jlayton@kernel.org> says:
The VFS has always used coarse-grained timestamps when updating the
ctime and mtime after a change. This has the benefit of allowing
filesystems to optimize away a lot metadata updates, down to around 1
per jiffy, even when a file is under heavy writes.
Unfortunately, this has always been an issue when we're exporting via
NFSv3, which relies on timestamps to validate caches. A lot of changes
can happen in a jiffy, so timestamps aren't sufficient to help the
client decide when to invalidate the cache. Even with NFSv4, a lot of
exported filesystems don't properly support a change attribute and are
subject to the same problems with timestamp granularity. Other
applications have similar issues with timestamps (e.g backup
applications).
If we were to always use fine-grained timestamps, that would improve the
situation, but that becomes rather expensive, as the underlying
filesystem would have to log a lot more metadata updates.
What we need is a way to only use fine-grained timestamps when they are
being actively queried. Use the (unused) top bit in inode->i_ctime_nsec
as a flag that indicates whether the current timestamps have been
queried via stat() or the like. When it's set, we allow the kernel to
use a fine-grained timestamp iff it's necessary to make the ctime show
a different value.
This solves the problem of being able to distinguish the timestamp
between updates, but introduces a new problem: it's now possible for a
file being changed to get a fine-grained timestamp. A file that is
altered just a bit later can then get a coarse-grained one that appears
older than the earlier fine-grained time. This violates timestamp
ordering guarantees.
To remedy this, keep a global monotonic atomic64_t value that acts as a
timestamp floor. When we go to stamp a file, we first get the latter of
the current floor value and the current coarse-grained time. If the
inode ctime hasn't been queried then we just attempt to stamp it with
that value.
If it has been queried, then first see whether the current coarse time
is later than the existing ctime. If it is, then we accept that value.
If it isn't, then we get a fine-grained time and try to swap that into
the global floor. Whether that succeeds or fails, we take the resulting
floor time, convert it to realtime and try to swap that into the ctime.
We take the result of the ctime swap whether it succeeds or fails, since
either is just as valid.
Filesystems can opt into this by setting the FS_MGTIME fstype flag.
Others should be unaffected (other than being subject to the same floor
value as multigrain filesystems).
* patches from https://lore.kernel.org/r/20241002-mgtime-v10-0-d1c4717f5284@kernel.org:
tmpfs: add support for multigrain timestamps
btrfs: convert to multigrain timestamps
ext4: switch to multigrain timestamps
xfs: switch to multigrain timestamps
Documentation: add a new file documenting multigrain timestamps
fs: add percpu counters for significant multigrain timestamp events
fs: tracepoints around multigrain timestamp events
fs: handle delegated timestamps in setattr_copy_mgtime
fs: have setattr_copy handle multigrain timestamps appropriately
fs: add infrastructure for multigrain timestamps
Link: https://lore.kernel.org/r/20241002-mgtime-v10-0-d1c4717f5284@kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Enable multigrain timestamps, which should ensure that there is an
apparent change to the timestamp whenever it has been written after
being actively observed via getattr.
Also, anytime the mtime changes, the ctime must also change, and those
are now the only two options for xfs_trans_ichgtime. Have that function
unconditionally bump the ctime, and ASSERT that XFS_ICHGTIME_CHG is
always set.
Finally, stop setting STATX_CHANGE_COOKIE in getattr, since the ctime
should give us better semantics now.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org> # documentation bits
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Link: https://lore.kernel.org/r/20241002-mgtime-v10-9-d1c4717f5284@kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>
Currently the debug-only xfs_bmap_exact_minlen_extent_alloc allocation
variant fails to drop into the lowmode last resort allocator, and
thus can sometimes fail allocations for which the caller has a
transaction block reservation.
Fix this by using xfs_bmap_btalloc_low_space to do the actual allocation.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
xfs_bmap_exact_minlen_extent_alloc duplicates the args setup in
xfs_bmap_btalloc. Switch to call it from xfs_bmap_btalloc after
doing the basic setup.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Exact minlen allocations only exist as an error injection tool for debug
builds. Currently this is implemented using ifdefs, which means the code
isn't even compiled for non-XFS_DEBUG builds. Enhance the compile test
coverage by always building the code and use the compilers' dead code
elimination to remove it from the generated binary instead.
The only downside is that the alloc_minlen_only field is unconditionally
added to struct xfs_alloc_args now, but by moving it around and packing
it tightly this doesn't actually increase the size of the structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Userdata and metadata allocations end up in the same allocation helpers.
Remove the separate xfs_bmap_alloc_userdata function to make this more
clear.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
Just like xfs_attr3_leaf_split, xfs_attr_node_try_addname can return
-ENOSPC both for an actual failure to allocate a disk block, but also
to signal the caller to convert the format of the attr fork. Use magic
1 to ask for the conversion here as well.
Note that unlike the similar issue in xfs_attr3_leaf_split, this one was
only found by code review.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
xfs_attr3_leaf_split propagates the need for an extra btree split as
-ENOSPC to it's only caller, but the same return value can also be
returned from xfs_da_grow_inode when it fails to find free space.
Distinguish the two cases by returning 1 for the extra split case instead
of overloading -ENOSPC.
This can be triggered relatively easily with the pending realtime group
support and a file system with a lot of small zones that use metadata
space on the main device. In this case every about 5-10th run of
xfs/538 runs into the following assert:
ASSERT(oldblk->magic == XFS_ATTR_LEAF_MAGIC);
in xfs_attr3_leaf_split caused by an allocation failure. Note that
the allocation failure is caused by another bug that will be fixed
subsequently, but this commit at least sorts out the error handling.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
xfs_attr3_leaf_add only has two potential return values, indicating if the
entry could be added or not. Replace the errno return with a bool so that
ENOSPC from it can't easily be confused with a real ENOSPC.
Remove the return value from the xfs_attr3_leaf_add_work helper entirely,
as it always return 0.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
xfs_attr_leaf_try_add is only called by xfs_attr_leaf_addname, and
merging the two will simplify a following error handling fix.
To facilitate this move the remote block state save/restore helpers up in
the file so that they don't need forward declarations now.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Carlos Maiolino <cem@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRAhzRXHqcMeLMyaSiRxhvAZXjcogUCZuQEvwAKCRCRxhvAZXjc
ohg3APwJWQnqFlBddcRl4yrPJ/cgcYSYAOdHb+E+blomSwdxcwEAmwsnLPNQOtw2
rxKvQfZqhVT437bl7RpPPZrHGxwTng8=
=6v1r
-----END PGP SIGNATURE-----
Merge tag 'vfs-6.12.blocksize' of gitolite.kernel.org:pub/scm/linux/kernel/git/vfs/vfs
Pull vfs blocksize updates from Christian Brauner:
"This contains the vfs infrastructure as well as the xfs bits to enable
support for block sizes (bs) larger than page sizes (ps) plus a few
fixes to related infrastructure.
There has been efforts over the last 16 years to enable enable Large
Block Sizes (LBS), that is block sizes in filesystems where bs > page
size. Through these efforts we have learned that one of the main
blockers to supporting bs > ps in filesystems has been a way to
allocate pages that are at least the filesystem block size on the page
cache where bs > ps.
Thanks to various previous efforts it is possible to support bs > ps
in XFS with only a few changes in XFS itself. Most changes are to the
page cache to support minimum order folio support for the target block
size on the filesystem.
A motivation for Large Block Sizes today is to support high-capacity
(large amount of Terabytes) QLC SSDs where the internal Indirection
Unit (IU) are typically greater than 4k to help reduce DRAM and so in
turn cost and space. In practice this then allows different
architectures to use a base page size of 4k while still enabling
support for block sizes aligned to the larger IUs by relying on high
order folios on the page cache when needed.
It also allows to take advantage of the drive's support for atomics
larger than 4k with buffered IO support in Linux. As described this
year at LSFMM, supporting large atomics greater than 4k enables
databases to remove the need to rely on their own journaling, so they
can disable double buffered writes, which is a feature different cloud
providers are already enabling through custom storage solutions"
* tag 'vfs-6.12.blocksize' of gitolite.kernel.org:pub/scm/linux/kernel/git/vfs/vfs: (22 commits)
Documentation: iomap: fix a typo
iomap: remove the iomap_file_buffered_write_punch_delalloc return value
iomap: pass the iomap to the punch callback
iomap: pass flags to iomap_file_buffered_write_punch_delalloc
iomap: improve shared block detection in iomap_unshare_iter
iomap: handle a post-direct I/O invalidate race in iomap_write_delalloc_release
docs:filesystems: fix spelling and grammar mistakes in iomap design page
filemap: fix htmldoc warning for mapping_align_index()
iomap: make zero range flush conditional on unwritten mappings
iomap: fix handling of dirty folios over unwritten extents
iomap: add a private argument for iomap_file_buffered_write
iomap: remove set_memor_ro() on zero page
xfs: enable block size larger than page size support
xfs: make the calculation generic in xfs_sb_validate_fsb_count()
xfs: expose block size in stat
xfs: use kvmalloc for xattr buffers
iomap: fix iomap_dio_zero() for fs bs > system page size
filemap: cap PTE range to be created to allowed zero fill in folio_map_range()
mm: split a folio in minimum folio order chunks
readahead: allocate folios with mapping_min_order in readahead
...
COW writes remove the amount overwritten either directly for delalloc
reservations, or in earlier deferred transactions than adding the new
amount back in the bmap map transaction. This means st_blocks on an
inode where all data is overwritten using the COW path can temporarily
show a 0 st_blocks. This can easily be reproduced with the pending
zoned device support where all writes use this path and trips the
check in generic/615, but could also happen on a reflink file without
that.
Fix this by temporarily add the pending blocks to be mapped to
i_delayed_blks while the item is queued.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
Convert the perag lookup from the legacy radix tree to the xarray,
which allows for much nicer iteration and bulk lookup semantics.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
The tagged perag helpers are only used in xfs_icache.c in the kernel code
and not at all in xfsprogs. Move them to xfs_icache.c in preparation for
switching to an xarray, for which I have no plan to implement the tagged
lookup functions for userspace.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
Using the kfree_rcu_mightsleep is simpler and removes the need for a
rcu_head in the perag structure.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
./fs/xfs/libxfs/xfs_defer.c: xfs_trans_priv.h is included more than once.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=9491
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
We checked that "pip" is non-NULL at the start of the if else statement
so there is no need to check again here. Delete the check.
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
Pankaj Raghav reported that when filesystem block size is larger
than page size, the xattr code can use kmalloc() for high order
allocations. This triggers a useless warning in the allocator as it
is a __GFP_NOFAIL allocation here:
static inline
struct page *rmqueue(struct zone *preferred_zone,
struct zone *zone, unsigned int order,
gfp_t gfp_flags, unsigned int alloc_flags,
int migratetype)
{
struct page *page;
/*
* We most definitely don't want callers attempting to
* allocate greater than order-1 page units with __GFP_NOFAIL.
*/
>>>> WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
...
Fix this by changing all these call sites to use kvmalloc(), which
will strip the NOFAIL from the kmalloc attempt and if that fails
will do a __GFP_NOFAIL vmalloc().
This is not an issue that productions systems will see as
filesystems with block size > page size cannot be mounted by the
kernel; Pankaj is developing this functionality right now.
Reported-by: Pankaj Raghav <kernel@pankajraghav.com>
Fixes: f078d4ea82 ("xfs: convert kmem_alloc() to kmalloc()")
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Link: https://lore.kernel.org/r/20240822135018.1931258-8-kernel@pankajraghav.com
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Pankaj Raghav <p.raghav@samsung.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Daniel Gomez <da.gomez@samsung.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Standardize the parameters in xfs_{alloc,bm,ino,rmap,refcount}bt_maxrecs
so that we have consistent calling conventions. This doesn't affect the
kernel that much, but enables us to clean up userspace a bit.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
While refactoring code, I noticed that when xfs_iroot_realloc tries to
shrink a bmbt root block, it allocates a smaller new block and then
copies "records" and pointers to the new block. However, bmbt root
blocks cannot ever be leaves, which means that it's not technically
correct to copy records. We /should/ be copying keys.
Note that this has never resulted in actual memory corruption because
sizeof(bmbt_rec) == (sizeof(bmbt_key) + sizeof(bmbt_ptr)). However,
this will no longer be true when we start adding realtime rmap stuff,
so fix this now.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Several people reported C++ compilation errors due to things that C
compilers allow but C++ compilers do not. Fix both of these problems,
and hope there aren't more of these brown paper bags in 2 months when we
finally get these fixes through the process into a released xfsprogs.
NOTE: I am submitting this bugfix over the objections of a former
maintainer, who insists that we should remove this function from the
published userspace ABI instead of fixing the C++ compilation errors.
No deprecation period, no discussion, just a hard drop of an already
provided and correct C function, which would be in contravention of
Linus' rules. IOWs, removing ABI that have already shipped in a
released kernel requires a careful deprecation period, so I will let
that maintainer run that process.
Reported-by: kernel@mattwhitlock.name
Reported-by: sam@gentoo.org
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219203
Fixes: 233f4e12bb ("xfs: add parent pointer ioctls")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Track the RT summary file size in blocks, just like the RT bitmap
file. While we have users of both units, blocks are used slightly
more often and this matches the bitmap file for consistency.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
xfs_rtbitmap_wordcount and xfs_rtsummary_wordcount are currently unused,
so remove them to simplify refactoring other rtbitmap helpers. They
can be added back or simply open coded when actually needed.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Turn the ISVALID macro defined and used inside in xfs_bmap_adjacent
that relies on implict context into a proper inline function.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
There isn't much of a good reason to pass the xfs_rtalloc_rec structures
that describe extents to xfs_rtalloc_query_range as we really just want
a lower and upper bound xfs_rtxnum_t. Pass the rtxnum directly and
simply the interface.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Simplify the number of block number conversion helpers by removing
xfs_rtb_to_rtxrem. Any recent compiler is smart enough to eliminate
the double divisions if using separate xfs_rtb_to_rtx and
xfs_rtb_to_rtxoff calls.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
When growfs sets an extent size, it doesn't updated the m_rtxblklog and
m_rtxblkmask values, which could lead to incorrect usage of them if they
were set before and can't be used for the new extent size.
Add a xfs_mount_sb_set_rextsize helper that updates the two fields, and
also use it when calculating the new RT geometry instead of disabling
the optimization there.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
To prepare for being able to join an already locked rtbitmap inode to a
transaction split out separate helpers for joining the transaction from
the locking helpers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Add helpers to libxfs that can be shared by growfs and mkfs for
initializing the rtbitmap and summary, and by passing the optional data
pointer also by repair for rebuilding them. This will become even more
useful when the rtgroups feature adds a metadata header to each block,
which means even more shared code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
[djwong: minor documentation and data advance tweaks]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Add a corruption check for passing an invalid block number, which is a
lot easier to understand than the xfs_bmapi_read failure later on.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Protect against developers passing stupid limits when refactoring the
RT code once again.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
All callers pass a 0 limit to xfs_rtfind_back, so remove the argument
and hard code it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Split the RT geometry validation in the early mount code into a
helper than can be reused by repair (from which this code was
apparently originally stolen anyway).
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
[djwong: u64 return value for calc_rbmblocks]
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Replace xfs_validate_rtextents with an open coded check for 0
rtextents. The name for the function implies it does a lot more
than a zero check, which is more obvious when open coded.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Pass the xfs_icreate_args object to xfs_dialloc since we can extract the
relevant mode (really just the file type) and parent inumber from there.
This simplifies the calling convention in preparation for the next
patch.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
This patch introduces two more new ioctls to manage atomic updates to
file contents -- XFS_IOC_START_COMMIT and XFS_IOC_COMMIT_RANGE. The
commit mechanism here is exactly the same as what XFS_IOC_EXCHANGE_RANGE
does, but with the additional requirement that file2 cannot have changed
since some sampling point. The start-commit ioctl performs the sampling
of file attributes.
Note: This patch currently samples i_ctime during START_COMMIT and
checks that it hasn't changed during COMMIT_RANGE. This isn't entirely
safe in kernels prior to 6.12 because ctime only had coarse grained
granularity and very fast updates could collide with a COMMIT_RANGE.
With the multi-granularity ctime introduced by Jeff Layton, it's now
possible to update ctime such that this does not happen.
It is critical, then, that this patch must not be backported to any
kernel that does not support fine-grained file change timestamps.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Acked-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
As a result of the factoring in commit 14dd46cf31 ("xfs: split
xfs_inobt_init_cursor"), mount started taking a long time on a
user's filesystem. For Anders, this made mount times regress from
under a second to over 15 minutes for a filesystem with only 30
million inodes in it.
Anders bisected it down to the above commit, but even then the bug
was not obvious. In this commit, over 20 calls to
xfs_inobt_init_cursor() were modified, and some we modified to call
a new function named xfs_finobt_init_cursor().
If that takes you a moment to reread those function names to see
what the rename was, then you have realised why this bug wasn't
spotted during review. And it wasn't spotted on inspection even
after the bisect pointed at this commit - a single missing "f" isn't
the easiest thing for a human eye to notice....
The result is that xfs_finobt_count_blocks() now incorrectly calls
xfs_inobt_init_cursor() so it is now walking the inobt instead of
the finobt. Hence when there are lots of allocated inodes in a
filesystem, mount takes a -long- time run because it now walks a
massive allocated inode btrees instead of the small, nearly empty
free inode btrees. It also means all the finobt space reservations
are wrong, so mount could potentially given ENOSPC on kernel
upgrade.
In hindsight, commit 14dd46cf31 should have been two commits - the
first to convert the finobt callers to the new API, the second to
modify the xfs_inobt_init_cursor() API for the inobt callers. That
would have made the bug very obvious during review.
Fixes: 14dd46cf31 ("xfs: split xfs_inobt_init_cursor")
Reported-by: Anders Blomdell <anders.blomdell@gmail.com>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
"KjellR" complained on IRC that an old V4 filesystem suddenly stopped
mounting after upgrading from 6.9.11 to 6.10.3, with the following splat
when trying to read the rt bitmap inode:
00000000: 49 4e 80 00 01 02 00 01 00 00 00 00 00 00 00 00 IN..............
00000010: 00 00 00 01 00 00 00 00 00 00 00 00 00 00 00 00 ................
00000020: 00 00 00 00 00 00 00 00 43 d2 a9 da 21 0f d6 30 ........C...!..0
00000030: 43 d2 a9 da 21 0f d6 30 00 00 00 00 00 00 00 00 C...!..0........
00000040: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00000050: 00 00 00 02 00 00 00 00 00 00 00 04 00 00 00 00 ................
00000060: ff ff ff ff 00 00 00 00 00 00 00 00 00 00 00 00 ................
00000070: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
As Dave Chinner points out, this is a V1 inode with both di_onlink and
di_nlink set to 1 and di_flushiter == 0. In other words, this inode was
formatted this way by mkfs and hasn't been touched since then.
Back in the old days of xfsprogs 3.2.3, I observed that libxfs_ialloc
would set di_nlink, but if the filesystem didn't have NLINK, it would
then set di_version = 1. libxfs_iflush_int later sees the V1 inode and
copies the value of di_nlink to di_onlink without zeroing di_onlink.
Eventually this filesystem must have been upgraded to support NLINK
because 6.10 doesn't support !NLINK filesystems, which is how we tripped
over this old behavior. The filesystem doesn't have a realtime section,
so that's why the rtbitmap inode has never been touched.
Fix this by removing the di_onlink/di_nlink checking for all V1/V2
inodes because this is a muddy mess. The V3 inode handling code has
always supported NLINK and written di_onlink==0 so keep that check.
The removal of the V1 inode handling code when we dropped support for
!NLINK obscured this old behavior.
Reported-by: kjell.m.randa@gmail.com
Fixes: 40cb8613d6 ("xfs: check unused nlink fields in the ondisk inode")
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
In the macro definition of XFS_DQUOT_LOGRES, a parameter is accepted,
but it is not used. Hence, it should be removed.
This patch has only passed compilation test, but it should be fine.
Signed-off-by: Julian Sun <sunjunchao2870@gmail.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
The pag in xfs_ag_resv_rmapbt_alloc() is already held when the struct
xfs_btree_cur is initialized in xfs_rmapbt_init_cursor(), so there is no
need to get pag again.
On the other hand, in xfs_rmapbt_free_block(), the similar function
xfs_ag_resv_rmapbt_free() was removed in commit 92a005448f ("xfs: get
rid of unnecessary xfs_perag_{get,put} pairs"), xfs_ag_resv_rmapbt_alloc()
was left because scrub used it, but now scrub has removed it. Therefore,
we could get rid of xfs_ag_resv_rmapbt_alloc() just like the rmap free
block, make the code cleaner.
Signed-off-by: Long Li <leo.lilong@huawei.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
Currently the AIL attempts to keep 25% of the "log space" free,
where the current used space is tracked by the reserve grant head.
That is, it tracks both physical space used plus the amount reserved
by transactions in progress.
When we start tail pushing, we are trying to make space for new
reservations by writing back older metadata and the log is generally
physically full of dirty metadata, and reservations for modifications
in flight take up whatever space the AIL can physically free up.
Hence we don't really need to take into account the reservation
space that has been used - we just need to keep the log tail moving
as fast as we can to free up space for more reservations to be made.
We know exactly how much physical space the journal is consuming in
the AIL (i.e. max LSN - min LSN) so we can base push thresholds
directly on this state rather than have to look at grant head
reservations to determine how much to physically push out of the
log.
This also allows code that needs to know if log items in the current
transaction need to be pushed or re-logged to simply sample the
current target - they don't need to calculate the current target
themselves. This avoids the need for any locking when doing such
checks.
Further, moving to a physical target means we don't need "push all
until empty semantics" like were introduced in the previous patch.
We can now test and clear the "push all" as a one-shot command to
set the target to the current head of the AIL. This allows the
xfsaild to maximise the use of log space right up to the point where
conditions indicate that the xfsaild is not keeping up with load and
it needs to work harder, and as soon as those constraints go away
(i.e. external code no longer needs everything pushed) the xfsaild
will return to maintaining the normal 25% free space thresholds.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
We have a mechanism that checks the amount of log space remaining
available every time we make a transaction reservation. If the
amount of space is below a threshold (25% free) we push on the AIL
to tell it to do more work. To do this, we end up calculating the
LSN that the AIL needs to push to on every reservation and updating
the push target for the AIL with that new target LSN.
This is silly and expensive. The AIL is perfectly capable of
calculating the push target itself, and it will always be running
when the AIL contains objects.
What the target does is determine if the AIL needs to do
any work before it goes back to sleep. If we haven't run out of
reservation space or memory (or some other push all trigger), it
will simply go back to sleep for a while if there is more than 25%
of the journal space free without doing anything.
If there are items in the AIL at a lower LSN than the target, it
will try to push up to the target or to the point of getting stuck
before going back to sleep and trying again soon after.`
Hence we can modify the AIL to calculate it's own 25% push target
before it starts a push using the same reserve grant head based
calculation as is currently used, and remove all the places where we
ask the AIL to push to a new 25% free target. We can also drop the
minimum free space size of 256BBs from the calculation because the
25% of a minimum sized log is *always going to be larger than
256BBs.
This does still require a manual push in certain circumstances.
These circumstances arise when the AIL is not full, but the
reservation grants consume the entire of the free space in the log.
In this case, we still need to push on the AIL to free up space, so
when we hit this condition (i.e. reservation going to sleep to wait
on log space) we do a single push to tell the AIL it should empty
itself. This will keep the AIL moving as new reservations come in
and want more space, rather than keep queuing them and having to
push the AIL repeatedly.
The reason for using the "push all" when grant space runs out is
that we can run out of grant space when there is more than 25% of
the log free. Small logs are notorious for this, and we have a hack
in the log callback code (xlog_state_set_callback()) where we push
the AIL because the *head* moved) to ensure that we kick the AIL
when we consume space in it because that can push us over the "less
than 25% available" available that starts tail pushing back up
again.
Hence when we run out of grant space and are going to sleep, we have
to consider that the grant space may be consuming almost all the log
space and there is almost nothing in the AIL. In this situation, the
AIL pins the tail and moving the tail forwards is the only way the
grant space will come available, so we have to force the AIL to push
everything to guarantee grant space will eventually be returned.
Hence triggering a "push all" just before sleeping removes all the
nasty corner cases we have in other parts of the code that work
around the "we didn't ask the AIL to push enough to free grant
space" condition that leads to log space hangs...
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
A concurrent file creation and little writing could unexpectedly return
-ENOSPC error since there is a race window that the allocator could get
the wrong agf->agf_longest.
Write file process steps:
1) Find the entry that best meets the conditions, then calculate the start
address and length of the remaining part of the entry after allocation.
2) Delete this entry and update the -current- agf->agf_longest.
3) Insert the remaining unused parts of this entry based on the
calculations in 1), and update the agf->agf_longest again if necessary.
Create file process steps:
1) Check whether there are free inodes in the inode chunk.
2) If there is no free inode, check whether there has space for creating
inode chunks, perform the no-lock judgment first.
3) If the judgment succeeds, the judgment is performed again with agf lock
held. Otherwire, an error is returned directly.
If the write process is in step 2) but not go to 3) yet, the create file
process goes to 2) at this time, it may be mistaken for no space,
resulting in the file system still has space but the file creation fails.
We have sent two different commits to the community in order to fix this
problem[1][2]. Unfortunately, both solutions have flaws. In [2], I
discussed with Dave and Darrick, realized that a better solution to this
problem requires the "last cnt record tracking" to be ripped out of the
generic btree code. And surprisingly, Dave directly provided his fix code.
This patch includes appropriate modifications based on his tmp-code to
address this issue.
The entire fix can be roughly divided into two parts:
1) Delete the code related to lastrec-update in the generic btree code.
2) Place the process of updating longest freespace with cntbt separately
to the end of the cntbt modifications. Move the cursor to the rightmost
firstly, and update the longest free extent based on the record.
Note that we can not update the longest with xfs_alloc_get_rec() after
find the longest record, as xfs_verify_agbno() may not pass because
pag->block_count is updated on the outside. Therefore, use
xfs_btree_get_rec() as a replacement.
[1] https://lore.kernel.org/all/20240419061848.1032366-2-yebin10@huawei.com
[2] https://lore.kernel.org/all/20240604071121.3981686-1-wozizhi@huawei.com
Reported by: Ye Bin <yebin10@huawei.com>
Signed-off-by: Zizhi Wo <wozizhi@huawei.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
