We've had rather rare reports of bmap btree block corruption where
the bmap root block has a level count of zero. The root cause of the
corruption is so far unknown. We do have verifier checks to detect
this form of on-disk corruption, but this doesn't cover a memory
corruption variant of the problem. The latter is a reasonable
possibility because the root block is part of the inode fork and can
reside in-core for some time before inode extents are read.

If this occurs, it leads to a system crash such as the following:

 BUG: unable to handle kernel paging request at ffffffff00000221
 PF error: [normal kernel read fault]
 ...
 RIP: 0010:xfs_trans_brelse+0xf/0x200 [xfs]
 ...
 Call Trace:
  xfs_iread_extents+0x379/0x540 [xfs]
  xfs_file_iomap_begin_delay+0x11a/0xb40 [xfs]
  ? xfs_attr_get+0xd1/0x120 [xfs]
  ? iomap_write_begin.constprop.40+0x2d0/0x2d0
  xfs_file_iomap_begin+0x4c4/0x6d0 [xfs]
  ? __vfs_getxattr+0x53/0x70
  ? iomap_write_begin.constprop.40+0x2d0/0x2d0
  iomap_apply+0x63/0x130
  ? iomap_write_begin.constprop.40+0x2d0/0x2d0
  iomap_file_buffered_write+0x62/0x90
  ? iomap_write_begin.constprop.40+0x2d0/0x2d0
  xfs_file_buffered_aio_write+0xe4/0x3b0 [xfs]
  __vfs_write+0x150/0x1b0
  vfs_write+0xba/0x1c0
  ksys_pwrite64+0x64/0xa0
  do_syscall_64+0x5a/0x1d0
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

The crash occurs because xfs_iread_extents() attempts to release an
uninitialized buffer pointer as the level == 0 value prevented the
buffer from ever being allocated or read. Change the level > 0
assert to an explicit error check in xfs_iread_extents() to avoid
crashing the kernel in the event of localized, in-core inode
corruption.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Remove typedefs and consolidate local variable initialization.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NAllison Henderson <allison.henderson@oracle.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>

Smatch complains about the following:

fs/xfs/libxfs/xfs_dir2_leaf.c:848 xfs_dir2_leaf_addname() error:
uninitialized symbol 'lowstale'.

fs/xfs/libxfs/xfs_dir2_leaf.c:849 xfs_dir2_leaf_addname() error:
uninitialized symbol 'highstale'.

I don't think there's any incorrect behavior associated with the
uninitialized variable, but as the author of the previous zero-init
patch points out, it's best not to be passing around pointers to
uninitialized stack areas.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NNathan Chancellor <natechancellor@gmail.com>
Reviewed-by: NAllison Henderson <allison.henderson@oracle.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>

Remove typedefs and consolidate local variable initialization.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NNick Desaulniers <ndesaulniers@google.com>

When building with -Wsometimes-uninitialized, Clang warns:

fs/xfs/libxfs/xfs_dir2_node.c:481:6: warning: variable 'lowstale' is
used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
fs/xfs/libxfs/xfs_dir2_node.c:481:6: warning: variable 'highstale' is
used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]

While it isn't technically wrong, it isn't a problem in practice because
highstale and lowstale are only initialized in xfs_dir2_leafn_add when
compact is not zero then they are passed to xfs_dir3_leaf_find_entry,
where they are initialized before use when compact is zero. Regardless,
it's better not to be passing around uninitialized stack memory so zero
initialize these variables, which silences this warning.

Link: https://github.com/ClangBuiltLinux/linux/issues/393Signed-off-by: NNathan Chancellor <natechancellor@gmail.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

smatch complained about some uninitialized error returns, so fix those.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NAllison Henderson <allison.henderson@oracle.com>

If we have racing buffered and direct I/O COW fork extents under
writeback can have been moved to the data fork by the time we call
xfs_reflink_convert_cow from xfs_submit_ioend.  This would be mostly
harmless as the block numbers don't change by this move, except for
the fact that xfs_bmapi_write will crash or trigger asserts when
not finding existing extents, even despite trying to paper over this
with the XFS_BMAPI_CONVERT_ONLY flag.

Instead of special casing non-transaction conversions in the already
way too complicated xfs_bmapi_write just add a new helper for the much
simpler non-transactional COW fork case, which simplify ignores not
found extents.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Create a separate magic16 check function so that we don't run afoul of
static checkers.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

This way we can actually count how many bytes got converted and how many
calls we need, unlike in the caller which doesn't have the detailed
view.

Note that this includes a slight change in behavior as the
xs_xstrat_quick is now bumped for every allocation instead of just the
one covering the requested writeback offset, which makes a lot more
sense.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

No need to deal with the transaction and the inode locking in the
caller. Note that we also switch to passing whichfork as the second
paramter, matching what most related functions do.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Delalloc conversion has traditionally been part of our function to
allocate blocks on disk (first xfs_bmapi, then xfs_bmapi_write), but
delalloc conversion is a little special as we really do not want
to allocate blocks over holes, for which we don't have reservations.

Split the delalloc conversions into a separate helper to keep the
code simple and structured.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

We want to be able to reuse them for the upcoming dedidcated delalloc
convert routine.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Move boilerplate code from the callers into xfs_bmap_btree_to_extents:

 - exit early without failure if we don't need to convert to the
   extent format
 - assert that we have a btree cursor
 - don't reinitialize the passed in logflags argument
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Rename this flag variable to imply more strongly that it's related to
the free inode btree (finobt) operation.  No functional changes.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Add dquot magic numbers to the buffer ops type, in case we ever want to
use them.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Use xfs_verify_magic to check the magic numbers of inodes.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

With the verifier magic value helper in place, we've left a bit more
duplicate code across the verifiers that involve struct
xfs_da3_blkinfo. This includes the da node, xattr leaf and dir leaf
verifiers, all of which perform similar checks for v4 and v5
filesystems.

Create a common helper to verify an xfs_da3_blkinfo structure,
taking care to only access v5 fields where appropriate, and refactor
the aforementioned verifiers to use the helper. No functional
changes.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Most buffer verifiers have hardcoded magic value checks
conditionalized on the version of the filesystem. The magic value
field of the verifier structure facilitates abstraction of some of
this code. Populate the ->magic field of various verifiers to take
advantage of this abstraction. No functional changes.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The dir2 leaf verifiers share the same underlying structure
verification code, but implement six accessor functions to multiplex
the code across the two verifiers. Further, the magic value isn't
sufficiently abstracted such that the common helper has to manually
fix up the magic from the caller on v5 filesystems.

Use the magic field in the verifier structure to eliminate the
duplicate code and clean this all up. No functional change.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The allocation btree verifiers share code that is unable to detect
cross-tree magic value corruptions such as a bnobt block with a
cntbt magic value. Populate the b_ops->magic field of the associated
verifier structures such that the structure verifier can check the
magic value against the expected value based on tree type.

The btree level check requires knowledge of the tree type to
determine the appropriate maximum value. This was previously part of
the hardcoded magic value checks. With that code removed, peek at
the first magic value in the verifier to determine the expected tree
type of the current block.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Similar to the inode btree verifier, the same allocation btree
verifier structure is shared between the by-bno (bnobt) and by-size
(cntbt) btrees. This prevents the ability to distinguish magic
values between them. Separate the verifier into two, one for each
tree, and assign them appropriately. No functional changes.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The inode btree verifier code is shared between the inode btree and
free inode btree because the underlying metadata formats are
essentially equivalent. A side effect of this is that the verifier
cannot determine whether a particular btree block should have an
inobt or finobt magic value.

This logic allows an unfortunate xfs_repair bug to escape detection
where certain level > 0 nodes of the finobt are stamped with inobt
magic by xfs_repair finobt reconstruction. This is fortunately not a
severe problem since the inode btree magic values do not contribute
to any changes in kernel behavior, but we do need a means to detect
and prevent this problem in the future.

Add a field to xfs_buf_ops to store the v4 and v5 superblock magic
values expected by a particular verifier. Add a helper to check an
on-disk magic value against the value expected by the verifier. Call
the helper from the shared [f]inobt verifier code for magic value
verification. This ensures that the inode btree blocks each have the
appropriate magic value based on specific tree type and superblock
version.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The inobt verifier is reused for the inobt and finobt, which
prevents the ability to distinguish between magic values on a
per-tree basis. Create a separate finobt structure in preparation
for changes to enforce the appropriate magic value for the
associated tree. This patch has no functional change.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Most verifiers that check on-disk magic values convert the CPU
endian magic value constant to disk endian to facilitate compile
time optimization of the byte swap and reduce the need for runtime
byte swaps in buffer verifiers. Several buffer verifiers do not
follow this pattern. Update those verifiers for consistency.

Also fix up a random typo in the inode readahead verifier name.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Use a rhashtable to cache the unlinked list incore.  This should speed
up unlinked processing considerably when there are a lot of inodes on
the unlinked list because iunlink_remove no longer has to traverse an
entire bucket list to find which inode points to the one being removed.

The incore list structure records "X.next_unlinked = Y" relations, with
the rhashtable using Y to index the records.  This makes finding the
inode X that points to a inode Y very quick.  If our cache fails to find
anything we can always fall back on the old method.

FWIW this drastically reduces the amount of time it takes to remove
inodes from the unlinked list.  I wrote a program to open a lot of
O_TMPFILE files and then close them in the same order, which takes
a very long time if we have to traverse the unlinked lists.  With the
ptach, I see:

+ /d/t/tmpfile/tmpfile
Opened 193531 files in 6.33s.
Closed 193531 files in 5.86s

real    0m12.192s
user    0m0.064s
sys     0m11.619s
+ cd /
+ umount /mnt

real    0m0.050s
user    0m0.004s
sys     0m0.030s

And without the patch:

+ /d/t/tmpfile/tmpfile
Opened 193588 files in 6.35s.
Closed 193588 files in 751.61s

real    12m38.853s
user    0m0.084s
sys     12m34.470s
+ cd /
+ umount /mnt

real    0m0.086s
user    0m0.000s
sys     0m0.060s
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Add a new helper to check that a per-AG inode pointer is either null or
points somewhere valid within that AG.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

The writeback delalloc conversion code is racy with respect to
changes in the currently cached file mapping outside of the current
page. This is because the ilock is cycled between the time the
caller originally looked up the mapping and across each real
allocation of the provided file range. This code has collected
various hacks over the years to help combat the symptoms of these
races (i.e., truncate race detection, allocation into hole
detection, etc.), but none address the fundamental problem that the
imap may not be valid at allocation time.

Rather than continue to use race detection hacks, update writeback
delalloc conversion to a model that explicitly converts the delalloc
extent backing the current file offset being processed. The current
file offset is the only block we can trust to remain once the ilock
is dropped because any operation that can remove the block
(truncate, hole punch, etc.) must flush and discard pagecache pages
first.

Modify xfs_iomap_write_allocate() to use the xfs_bmapi_delalloc()
mechanism to request allocation of the entire delalloc extent
backing the current offset instead of assuming the extent passed by
the caller is unchanged. Record the range specified by the caller
and apply it to the resulting allocated extent so previous checks by
the caller for COW fork overlap are not lost. Finally, overload the
bmapi delalloc flag with the range reval flag behavior since this is
the only use case for both.

This ensures that writeback always picks up the correct
and current extent associated with the page, regardless of races
with other extent modifying operations. If operating on a data fork
and the COW overlap state has changed since the ilock was cycled,
the caller revalidates against the COW fork sequence number before
using the imap for the next block.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The writeback delalloc conversion code is racy with respect to
changes in the currently cached file mapping. This stems from the
fact that the bmapi allocation code requires a file range to
allocate and the writeback conversion code assumes the range of the
currently cached mapping is still valid with respect to the fork. It
may not be valid, however, because the ilock is cycled (potentially
multiple times) between the time the cached mapping was populated
and the delalloc conversion occurs.

To facilitate a solution to this problem, create a new
xfs_bmapi_delalloc() wrapper to xfs_bmapi_write() that takes a file
(FSB) offset and attempts to allocate whatever delalloc extent backs
the offset. Use a new bmapi flag to cause xfs_bmapi_write() to set
the range based on the extent backing the bno parameter unless bno
lands in a hole. If bno does land in a hole, fall back to the
current behavior (which may result in an error or quietly skipping
holes in the specified range depending on other parameters). This
patch does not change behavior.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Now that the cached writeback mapping is explicitly invalidated on
data fork changes, the EOF trimming band-aid is no longer necessary.
Remove xfs_trim_extent_eof() as well since it has no other users.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The sequence counter in the xfs_ifork structure is only updated on
COW forks. This is because the counter is currently only used to
optimize out repetitive COW fork checks at writeback time.

Tweak the extent code to update the seq counter regardless of the
fork type in preparation for using this counter on data forks as
well.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NAllison Henderson <allison.henderson@oracle.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Check extended attribute entry names for invalid characters.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Check directory entry names for invalid characters.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Teach scrub to flag extent maps that exceed the range that can be mapped
with a xfs_dablk_t.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Use __print_symbolic to print the scrub type in ftrace output.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

Use __print_symbolic to print the btree type in ftrace output.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

Move XFS_INODE_FORMAT_STR to libxfs so that we don't forget to keep it
updated, and add necessary TRACE_DEFINE_ENUM.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

Move XFS_AG_BTREE_CMP_FORMAT_STR to libxfs so that we don't forget to
keep it updated, and TRACE_DEFINE_ENUM the values while we're at it.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

ftrace's __print_symbolic() has a (very poorly documented) requirement
that any enum values used in the symbol to string translation table be
wrapped in a TRACE_DEFINE_ENUM so that the enum value can be encoded in
the ftrace ring buffer.  Fix this unsatisfied requirement.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>

The realtime summary is a two-dimensional array on disk, effectively:

u32 rsum[log2(number of realtime extents) + 1][number of blocks in the bitmap]

rsum[log][bbno] is the number of extents of size 2**log which start in
bitmap block bbno.

xfs_rtallocate_extent_near() uses xfs_rtany_summary() to check whether
rsum[log][bbno] != 0 for any log level. However, the summary array is
stored in row-major order (i.e., like an array in C), so all of these
entries are not adjacent, but rather spread across the entire summary
file. In the worst case (a full bitmap block), xfs_rtany_summary() has
to check every level.

This means that on a moderately-used realtime device, an allocation will
waste a lot of time finding, reading, and releasing buffers for the
realtime summary. In particular, one of our storage services (which runs
on servers with 8 very slow CPUs and 15 8 TB XFS realtime filesystems)
spends almost 5% of its CPU cycles in xfs_rtbuf_get() and
xfs_trans_brelse() called from xfs_rtany_summary().

One solution would be to also store the summary with the dimensions
swapped. However, this would require a disk format change to a very old
component of XFS.

Instead, we can cache the minimum size which contains any extents. We do
so lazily; rather than guaranteeing that the cache contains the precise
minimum, it always contains a loose lower bound which we tighten when we
read or update a summary block. This only uses a few kilobytes of memory
and is already serialized via the realtime bitmap and summary inode
locks, so the cost is minimal. With this change, the same workload only
spends 0.2% of its CPU cycles in the realtime allocator.
Signed-off-by: NOmar Sandoval <osandov@fb.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Store the inode cluster alignment information in units of inodes and
blocks in the mount data so that we don't have to keep recalculating
them.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>