Chris Dunlop reports a problem where an xattr operation fails,
reports the following error to syslog and hangs during unmount:

 ================================================
 [ BUG: lock held when returning to user space! ]
 ...
 ------------------------------------------------
 <PID> is leaving the kernel with locks still held!
 1 lock held by <PID>:
  #0:  (sb_internal){......}, at: [<ffffffffa07692a3>] xfs_trans_alloc+0xe3/0x130 [xfs]

The failure/shutdown occurs during deferred ops processing which
leads to an error return from xfs_defer_finish() via
xfs_attr_leaf_addname(). While the root cause of the failure is
unknown corruption, the cause of the subsequent BUG above and
unmount hang is failure to cancel the transaction before returning
to userspace.

The transaction is not cancelled because the out_defer_cancel error
handling paths in the xfs_attr_[leaf|node]_[add|remove]name()
functions clear args.trans without releasing the transaction. The
callers therefore lose the reference to the transaction and fail to
cancel it.

Since xfs_attr_[set|remove]() always cancel args.trans when != NULL
and xfs_defer_finish()->...->xfs_trans_roll() should always return
with a valid transaction, update the leaf/node xattr functions to
not reset args.trans in the error path responsible for cancelling
deferred ops.
Reported-by: NChris Dunlop <chris@onthe.net.au>
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>

XFS started using the perag metadata reservation pool for free inode
btree blocks in commit 76d771b4 ("xfs: use per-AG reservations
for the finobt"). To handle backwards compatibility, finobt blocks
are accounted against the pool so long as the full reservation is
available at mount time. Otherwise the ->m_inotbt_nores flag is set
and the filesystem falls back to the traditional per-transaction
finobt reservation.

This commit has two problems:

- finobt blocks are always accounted against the metadata
  reservation on allocation, regardless of ->m_inotbt_nores state
- finobt blocks are never returned to the reservation pool on free

The first problem affects reflink+finobt filesystems where the full
finobt reservation is not available at mount time. finobt blocks are
essentially stolen from the reflink reservation, putting refcountbt
management at risk of allocation failure. The second problem is an
unconditional leak of metadata reservation whenever finobt is
enabled.

Update the finobt block allocation callouts to consider
->m_inotbt_nores and account blocks appropriately. Blocks should be
consistently accounted against the metadata pool when
->m_inotbt_nores is false and otherwise tagged as RESV_NONE.
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>

It appears that the check for versions 4 or more is incorrect and is
off-by-one. Fix this.

Detected by CoverityScan, CID#1463775 ("Logically dead code")

Fixes: ac503a4c ("xfs: refactor the geometry structure filling function")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The mutex pag_ici_reclaim_lock of xfs_perag_t structure is initialized in
xfs_initialize_perag. If happen errors in xfs_initialize_perag, or free
resources in xfs_free_perag, wo need to destroy the mutex before free
perag.
Signed-off-by: NXiongwei Song <sxwjean@me.com>
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>

Starting with commit 57e73442 ("vsprintf: refactor %pK code out of
pointer"), the behavior of the raw '%p' printk format specifier was
changed to print a 32-bit hash of the pointer value to avoid leaking
kernel pointers into dmesg.  For most situations that's good.

This is /undesirable/ behavior when we're trying to debug XFS, however,
so define a PTR_FMT that prints the actual pointer when we're in debug
mode.

Note that %p for tracepoints still prints the raw pointer, so in the
long run we could consider rewriting some of these messages as
tracepoints.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Use the %pS instead of the %pF printk format specifier for printing
symbols from direct addresses. This is needed for the ia64, ppc64 and
parisc64 architectures.

While we're at it, be consistent with the capitalization of the 'S'.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Since %p prepends "0x" to the outputted string, we can drop the prefix.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

If a metadata IO error happens, we report the location of the failed IO
request in units of daddrs.  However, the printk message misleads people
into thinking that the units are fs blocks, so fix the reported units.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

If a malicious filesystem image contains a block+ format directory
wherein the directory inode's core.mode is set such that
S_ISDIR(core.mode) == 0, and if there are subdirectories of the
corrupted directory, an attempt to traverse up the directory tree will
crash the kernel in __xfs_dir3_data_check.  Running the online scrub's
parent checks will tend to do this.

The crash occurs because the directory inode's d_ops get set to
xfs_dir[23]_nondir_ops (it's not a directory) but the parent pointer
scrubber's indiscriminate call to xfs_readdir proceeds past the ASSERT
if we have non fatal asserts configured.

Fix the null pointer dereference crash in __xfs_dir3_data_check by
looking for S_ISDIR or wrong d_ops; and teach the parent scrubber
to bail out if it is fed a non-directory "parent".
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Refactor the geometry structure filling function to use the superblock
to fill the fields.  While we're at it, make the function less indenty
and use some whitespace to make the function easier to read.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Move xfs_fs_geometry to libxfs so that we can clean up the fs geometry
reporting in xfsprogs.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

At each mount, emit the transaction reservation type information via
tracepoints.  This makes it easier to compare the log reservation info
calculated by the kernel and xfsprogs so that we can more easily diagnose
minimum log size failures on freshly formatted filesystems.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Increase the corrupt buffer dump to the first 128 bytes since v5
filesystems have larger block headers than before.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Convert the two other error reporting functions to take xfs_failaddr_t
when the caller wishes to capture a code pointer instead of the classic
void * pointer.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Rename xfs_dqcheck to xfs_dquot_verify and make it return an
xfs_failaddr_t like every other structure verifier function.
This enables us to check on-disk quotas in the same way that we check
everything else.  Callers are now responsible for logging errors, as
XFS_QMOPT_DOWARN goes away.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Move the dquot repair code into a separate function and remove
XFS_QMOPT_DQREPAIR in favor of calling the helper directly.  Remove
other dead code because quotacheck is the only caller of DQREPAIR.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Expose all metadata structure buffer verifier functions via buf_ops.
These will be used by the online scrub mechanism to look for problems
with buffers that are already sitting around in memory.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

If the xattr leaf block looks corrupt, return -EFSCORRUPTED to userspace
instead of ASSERTing on debug kernels or running off the end of the
buffer on regular kernels.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Replace the current haphazard dir2 shortform verifier callsites with a
centralized verifier function that can be called either with the default
verifier functions or with a custom set.  This helps us strengthen
integrity checking while providing us with flexibility for repair tools.

xfs_repair wants this to be able to supply its own verifier functions
when trying to fix possibly corrupt metadata.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Change the short form directory structure verifier function to return
the instruction pointer of a failing check or NULL if everything's ok.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Create a function to check the structure of short form symlink targets.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Create a function to perform structure verification for short form
extended attributes.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Consolidate the fork size and format verifiers to xfs_dinode_verify so
that we can reject bad inodes earlier and in a single place.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Move the v3 inode integrity information (crc, owner, metauuid) before we
look at anything else in the inode so that we don't waste time on a torn
write or a totally garbled block.  This makes xfs_dinode_verify more
consistent with the other verifiers.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Refactor the callers of verifiers to print the instruction address of a
failing check.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Modify each function that checks the contents of a metadata buffer to
return the instruction address of the failing test so that we can report
more precise failure errors to the log.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Since all verification errors also mark the buffer as having an error,
we can combine these two calls.  Later we'll add a xfs_failaddr_t
parameter to promote the idea of reporting corruption errors and the
address of the failing check to enable better debugging reports.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Since __xfs_dir3_data_check verifies on-disk metadata, we can't have it
noisily blowing asserts and hanging the system on corrupt data coming in
off the disk.  Instead, have it return a boolean like all the other
checker functions, and only have it noisily fail if we fail in debug
mode.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Now that we have xfs_verify_agbno, use it to verify short form btree
pointers instead of open-coding them.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Create two helper functions to verify the headers of a long format
btree block.  We'll use this later for the realtime rmapbt.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

We already have a function to verify fsb pointers, so get rid of the
last users of the (less robust) macro.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

In xfs_scrub_get_inode, we don't do a good enough job distinguishing
EINVAL returns from xfs_iget w/ IGET_UNTRUSTED -- this can happen if the
passed in inode number is invalid (past eofs, inobt says it isn't an
inode) or if the inum is actually valid but the inode buffer fails
verifier.  In the first case we still want to return ENOENT, but in the
second case we want to capture the corruption error.

Therefore, if xfs_iget returns EINVAL, try the raw imap lookup.  If that
succeeds, we conclude it's a corruption error, otherwise we just bounce
out to userspace.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Always allocate a transaction for inode scrubbing, even if the _iget
fails.  This is something that is nice to have now for consistency with
the other scrubbers but will become critical when we get to online
repair where we'll actually use the transaction + raw buffer read to fix
the verifier errors.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Refactor xfs_scrub_bmap to use for_each_xfs_iext now that it exists.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

The superblock validation routines return a variety of error codes to
reject a mount request.  For scrub we can assume that the mount
succeeded, so if we see these things appear when scrubbing secondary sb
X, we can treat them all like corruption.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

In xfs_scrub_ag_read_headers, if we're not scrubbing the AGFL but
hit a read error reading the AGFL, we should reset the error code
so that it doesn't propagate up into the caller.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

If two programs simultaneously try to write to the same part of a file
via direct IO and buffered IO, there's a chance that the post-diowrite
pagecache invalidation will fail on the dirty page.  When this happens,
the dio write succeeded, which means that the page cache is no longer
coherent with the disk!

Programs are not supposed to mix IO types and this is a clear case of
data corruption, so store an EIO which will be reflected to userspace
during the next fsync.  Replace the WARN_ON with a ratelimited pr_crit
so that the developers have /some/ kind of breadcrumb to track down the
offending program(s) and file(s) involved.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NLiu Bo <bo.li.liu@oracle.com>

The create transaction reservation calculation has two different
branches of code depending on whether the filesystem is a v5 format
fs or older. Each branch considers the max reservation between the
allocation case (new chunk allocation + record insert) and the
modify case (chunk exists, record modification) of inode allocation.

The modify case is the same for both superblock versions with the
exception of the finobt. The finobt helper checks the feature bit,
however, and so the modify case already shares the same code.

Now that inode chunk allocation has been refactored into a helper
that checks the superblock version to calculate the appropriate
reservation for the create transaction, the only remaining
difference between the create and icreate branches is the call to
the finobt helper. As noted above, the finobt helper is a no-op when
the feature is not enabled. Therefore, these branches are
effectively duplicate and can be condensed.

Remove the xfs_calc_create_*() branch of functions and update the
various callers to use the xfs_calc_icreate_*() variant. The latter
creates the same reservation size for v4 create transactions as the
removed branch. As such, this patch does not result in transaction
reservation changes.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

The reservation for the various forms of inode allocation is
scattered across several different functions. This includes two
variants of chunk allocation (v5 icreate transactions vs. older
create transactions) and the inode free transaction.

To clean up some of this code and clarify the purpose of specific
allocfree reservations, continue the pattern of defining helper
functions for smaller operational units of broader transactions.
Refactor the reservation into an inode chunk alloc/free helper that
considers the various conditions based on filesystem format.

An inode chunk free involves an extent free and buffer
invalidations. The latter requires reservation for log headers only.
An inode chunk allocation modifies the free space btrees and logs
the chunk on v4 supers. v5 supers initialize the inode chunk using
ordered buffers and so do not log the chunk.

As a side effect of this refactoring, add one more allocfree res to
the ifree transaction. Technically this does not serve a specific
purpose because inode chunks are freed via deferred operations and
thus occur after a transaction roll. tr_ifree has a bit of a history
of tx overruns caused by too many agfl fixups during sustained file
deletion workloads, so add this extra reservation as a form of
padding nonetheless.
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Analysis of recent reports of log reservation overruns and code
inspection has uncovered that the reservations associated with inode
operations may not cover the worst case scenarios. In particular,
many cases only include one allocfree res. for a particular
operation even though said operations may also entail AGFL fixups
and inode btree block allocations in addition to the actual inode
chunk allocation. This can easily turn into two or three block
allocations (or frees) per operation.

In theory, the only way to define the worst case reservation is to
include an allocfree res for each individual allocation in a
transaction. Since that is impractical (we can perform multiple agfl
fixups per tx and not every allocation results in a full tree
operation), we need to find a reasonable compromise that addresses
the deficiency in practice without blowing out the size of the
transactions.

Since the inode btrees are not filled by the AGFL, record insertion
and removal can directly result in block allocations and frees
depending on the shape of the tree. These allocations and frees
occur in the same transaction context as the inobt update itself,
but are separate from the allocation/free that might be required for
an inode chunk. Therefore, it makes sense to assume that an [f]inobt
insert/remove can directly result in one or more block allocations
on behalf of the tree.

Refactor the inode transaction reservations to include one allocfree
res. per inode btree modification to cover allocations required by
the tree itself. This separates the reservation required to allocate
the inode chunk from the reservation required for inobt record
insertion/removal. Apply the same logic to the finobt. This results
in killing off the finobt modify condition because we no longer
assume that the broader transaction reservation will cover finobt
block allocations and finobt shape changes can occur in either of
the inobt allocation or modify situations.
Suggested-by: NDave Chinner <david@fromorbit.com>
Signed-off-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>