Even if we can't use the inobt/finobt cursors to count the number of
inode btree blocks, we are never allowed to clobber the cursor of the
btree being checked, so don't do this.  Found by fuzzing level = ones
in xfs/364.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Every so often we blow the ASSERT(type != XFS_IO_COW) in xfs_map_blocks
when running fsstress, as we do in generic/269.  The cause of this is
writeback racing with truncate -- writeback doesn't take the iolock, so
truncate can sneak in to decrease i_size and truncate page cache while
writeback is gathering buffer heads to schedule writeout.

If we hit this race on a block that has a CoW mapping, we'll get a valid
imap from the CoW fork but the reduced i_size trims the mapping to zero
length (which makes it invalid), so we call xfs_map_blocks to try again.
This doesn't do much anyway, since any mapping we get out of that will
also be invalid, so we might as well skip the assert and just stop.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Commit 66f36464 ("xfs: remove if_rdev") moved storing of rdev
value for special inodes to VFS inodes, but forgot to preserve the
value of i_rdev when recycling a reclaimable xfs_inode.

This was detected by xfstest overlay/017 with inodex=on mount option
and xfs base fs. The test does a lookup of overlay chardev and blockdev
right after drop caches.

Overlayfs inodes hold a reference on underlying xfs inodes when mount
option index=on is configured. If drop caches reclaim xfs inodes, before
it relclaims overlayfs inodes, that can sometimes leave a reclaimable xfs
inode and that test hits that case quite often.

When that happens, the xfs inode cache remains broken (zere i_rdev)
until the next cycle mount or drop caches.

Fixes: 66f36464 ("xfs: remove if_rdev")
Signed-off-by: NAmir Goldstein <amir73il@gmail.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>

Move all the inode and quota accounting updates out of xfs_bmap_btalloc
in preparation for fixing some quota accounting problems with copy on
write.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Refactor inode verifier error reporting into a non-libxfs function so
that we aren't encoding the message format in libxfs.  This also
changes the kernel dmesg output to resemble buffer verifier errors
more closely.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Fix all the inode number formats to be consistently (0x%llx) in all
trace point definitions.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Always zero the di_flags2 field when we free the inode so that we never
end up with an on-disk record for an unallocated inode that also has the
reflink iflag set.  This is in keeping with the general principle that
only files can have the reflink iflag set, even though we'll zero out
di_flags2 if we ever reallocate the inode.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Ensure that we've attached all the necessary dquots before performing
reflink operations so that quota accounting is accurate.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

Remove the extent size hint and realtime inode relevant code from
the xfs_bmapi_reserve_delalloc since it is not called on the inode
with extent size hint set or on a realtime inode.
Signed-off-by: NShan Hai <shan.hai@oracle.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Now that buffer's b_fspriv has been split, just replace the current
singly linked list of xfs_log_items, by the list_head infrastructure.

Also, remove the xfs_log_item argument from xfs_buf_resubmit_failed_buffers(),
there is no need for this argument, once the log items can be walked
through the list_head in the buffer.
Signed-off-by: NCarlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
[darrick: minor style cleanups]
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

By splitting the b_fspriv field into two different fields (b_log_item
and b_li_list). It's possible to get rid of an old ABI workaround, by
using the new b_log_item field to store xfs_buf_log_item separated from
the log items attached to the buffer, which will be linked in the new
b_li_list field.

This way, there is no more need to reorder the log items list to place
the buf_log_item at the beginning of the list, simplifying a bit the
logic to handle buffer IO.

This also opens the possibility to change buffer's log items list into a
proper list_head.

b_log_item field is still defined as a void *, because it is still used
by the log buffers to store xlog_in_core structures, and there is no
need to add an extra field on xfs_buf just for xlog_in_core.
Signed-off-by: NCarlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
[darrick: minor style changes]
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Take advantage of the rework on xfs_buf log items list, to get rid of
ths typedef for xfs_buf_log_item.

This patch also fix some indentation alignment issues found along the way.
Signed-off-by: NCarlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: NBill O'Donnell <billodo@redhat.com>
Reviewed-by: NDarrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>

Fix compiler warning on non-debug build
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Currently, we don't check sb_agblocks or sb_agblklog when we validate
the superblock, which means that we can fuzz garbage values into those
values and the mount succeeds.  This leads to all sorts of UBSAN
warnings in xfs/350 since we can then coerce other parts of xfs into
shifting by ridiculously large values.

Once we've validated agblocks, make sure the agcount makes sense.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Eryu Guan reported seeing occasional hangs when running generic/269 with
a new fsstress that supports clonerange/deduperange.  The cause of this
hang is an infinite loop when we convert the CoW fork extents from
unwritten to real just prior to writing the pages out; the infinite
loop happens because there's nothing in the CoW fork to convert, and so
it spins forever.

The fundamental issue here is that when we go to perform these CoW fork
conversions, we're supposed to have an extent waiting for us, but the
low space CoW reaper has snuck in and blown them away!  There are four
conditions that can dissuade the reaper from touching our file -- no
reflink iflag; dirty page cache; writeback in progress; or directio in
progress.  We check the four conditions prior to taking the locks, but
we neglect to recheck them once we have the locks, which is how we end
up whacking the writeback that's in progress.

Therefore, refactor the four checks into a helper function and call it
once again once we have the locks to make sure we really want to reap
the inode.  While we're at it, add an ASSERT for this weird condition so
that we'll fail noisily if we ever screw this up again.
Reported-by: NEryu Guan <eguan@redhat.com>
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Tested-by: NEryu Guan <eguan@redhat.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

xfs_bmbt_irec.br_blockcount is declared as xfs_filblks_t, which is an
unsigned 64-bit integer.  Though the bmbt helpers will never set a value
larger than 2^21 (since the underlying on-disk extent record has a
length field that is only 21 bits wide), we should be a little defensive
about checking that a bmbt record doesn't exceed what we're expecting or
overflow into the next AG.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

A btree format inode fork with zero records makes no sense, so reject it
if we see it, or else we can miscalculate memory allocations.  Found by
zeroes fuzzing {a,u3}.bmbt.numrecs in xfs/{374,378,412} with KASAN.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

In the attribute leaf verifier, we can check for obviously bad values of
firstused and count so that later attempts at lasthash don't run off the
end of the memory buffer.  Found by ones fuzzing hdr.count in xfs/400 with
KASAN.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

In xfs_scrub_dir_rec, we must walk through the directory block entries
to arrive at the offset given by the hash structure.  If we blindly
trust the hash address, we can end up midway into a directory entry and
stray outside the block.  Found by lastbit fuzzing lents[3].address in
xfs/390 with KASAN enabled.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Don't iunlock an unlocked inode, which can happen if the parent pointer
scrubber bails out with sc->ip unlocked while trying to grab the parent
directory inode.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NBrian Foster <bfoster@redhat.com>

Whenever we load a buffer, explicitly re-call the structure verifier to
ensure that memory isn't corrupting things.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Use an inode's block mappings to cross-reference inode block counters.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

While we're scrubbing various btrees, cross-reference the records
with the other metadata.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

During metadata btree scrub, we should cross-reference with the
reference counts.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Cross reference the refcount data with the rmap data to check that the
number of rmaps for a given block match the refcount of that block, and
that CoW blocks (which are owned entirely by the refcountbt) are tracked
as well.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

When scrubbing various btrees, we should cross-reference the records
with the reverse mapping btree and ensure that traversing the btree
finds the same number of blocks that the rmapbt thinks are owned by
that btree.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Cross-reference the inode btrees with the other metadata when we
scrub the filesystem.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Scrub should make sure that each bnobt record has a corresponding
cntbt record.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

When we're scrubbing various btrees, cross-reference the records with
the bnobt to ensure that we don't also think the space is free.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Create some stubs that will be used to cross-reference metadata records.
The actual cross-referencing will be filled in by subsequent patches.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

When scanning a metadata btree block, cross-reference the block location
with the free space btree and the reverse mapping btree to ensure that
the rmapbt knows about the block and the bnobt does not.  Add a
mechanism to defer checks when we happen to be scanning the bnobt/rmapbt
itself because it's less efficient to repeatedly clone and destroy the
cursor.

This patch provides the framework to make btree block owner checks
happen; the actual meat will be added in subsequent patches.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

There are a few places where we make a libxfs api call on behalf of some
object other than the one we're scrubbing but inadvertently call the
regular process_error function.  When this happens we mark the object
corrupt even though it was corruption in /some other/ object that
actually produced the -EFSCORRUPTED code.  The correct output flag for
these situations is SCRUB_OFLAG_XFAIL, not SCRUB_OFLAG_CORRUPT, so fix
this now that we also have a helper to set these.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Create some helper functions that we'll use later to deal with problems
we might encounter while cross referencing metadata with other metadata.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Add a couple of functions to the refcount btrees that will be used
to cross-reference metadata against the refcountbt.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Add a couple of functions to the rmap btrees that will be used
to cross-reference metadata against the rmapbt.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Add a couple of functions to the inode btrees that will be used
to cross-reference metadata against the inobt.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

Add a couple of functions to the free space btrees that will be used
to cross-reference metadata against the bnobt/cntbt, and a generic
btree function that provides the real implementation.
Signed-off-by: NDarrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: NDave Chinner <dchinner@redhat.com>

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>