Replace the current linear list and the indirection array for the in-core
extent list with a b+tree to avoid the need for larger memory allocations
for the indirection array when lots of extents are present. The current
extent list implementations leads to heavy pressure on the memory
allocator when modifying files with a high extent count, and can lead
to high latencies because of that.
The replacement is a b+tree with a few quirks. The leaf nodes directly
store the extent record in two u64 values. The encoding is a little bit
different from the existing in-core extent records so that the start
offset and length which are required for lookups can be retreived with
simple mask operations. The inner nodes store a 64-bit key containing
the start offset in the first half of the node, and the pointers to the
next lower level in the second half. In either case we walk the node
from the beginninig to the end and do a linear search, as that is more
efficient for the low number of cache lines touched during a search
(2 for the inner nodes, 4 for the leaf nodes) than a binary search.
We store termination markers (zero length for the leaf nodes, an
otherwise impossible high bit for the inner nodes) to terminate the key
list / records instead of storing a count to use the available cache
lines as efficiently as possible.
One quirk of the algorithm is that while we normally split a node half and
half like usual btree implementations we just spill over entries added at
the very end of the list to a new node on its own. This means we get a
100% fill grade for the common cases of bulk insertion when reading an
inode into memory, and when only sequentially appending to a file. The
downside is a slightly higher chance of splits on the first random
insertions.
Both insert and removal manually recurse into the lower levels, but
the bulk deletion of the whole tree is still implemented as a recursive
function call, although one limited by the overall depth and with very
little stack usage in every iteration.
For the first few extents we dynamically grow the list from a single
extent to the next powers of two until we have a first full leaf block
and that building the actual tree.
The code started out based on the generic lib/btree.c code from Joern
Engel based on earlier work from Peter Zijlstra, but has since been
rewritten beyond recognition.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[sandeen: use uint32_t for rec_len, update trace macros & repair/] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
To make life a little simpler make xfs_bmbt_set_all unaligned access
aware so that we can use it directly on the destination buffer.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Supporting a small bit of data inside the inode fork blows up the fork size
a lot, removing the 32 bytes of inline data halves the effective size of
the inode fork (and it still has a lot of unused padding left), and the
performance of a single kmalloc doesn't show up compared to the size to read
an inode or create one.
It also simplifies the fork management code a lot.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Instead of looking up extents to convert and calling xfs_bmapi_write on
each of them just let xfs_bmapi_write handle the full range. To make
this robust add a new XFS_BMAPI_CONVERT_ONLY that only converts ranges
and never allocates blocks.
[darrick: shorten the stringified CONVERT_ONLY trace flag]
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Add a new xfs_iext_cursor structure to hide the direct extent map
index manipulations. In addition to the existing lookup/get/insert/
remove and update routines new primitives to get the first and last
extent cursor, as well as moving up and down by one extent are
provided. Also new are convenience to increment/decrement the
cursor and retreive the new extent, as well as to peek into the
previous/next extent without updating the cursor and last but not
least a macro to iterate over all extents in a fork.
[darrick: rename for_each_iext to for_each_xfs_iext]
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[sandeen: use cursor in xfs_repair code as well] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This actually makes the function very slightly less efficient for now as we
detour through the expanded irect format between the in-core extent format
and the on-disk one instead of just endian swapping them. But with the
incore extent btree the in-core one will use a different format and the
representation will be entirely hidden.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This actually makes the function very slightly less efficient for now as we
detour through the expanded irect format between the in-core extent format
and the on-disk one instead of just endian swapping them. But with the
incore extent btree the in-core one will use a different format and the
representation will be entirely hidden. It also happens to make the
function a whole more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This prepares for getting rid of the current in-memory extent format.
At the end of the series we will change the calling convention again
to pass the xfs_bmbt_irec structure once it is available everywhere.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Stop poking before and after the index and just increment or decrement
it while doing our operations on it to prepare for a new extent list
implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Reported-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Two cases in xfs_bmap_add_extent_delay_real currently insert a new
extent before updating the existing one that is being split. While
this works fine with a simple extent list, a more complex tree can't
easily cope with overlapping extent. Reshuffle the code a bit to update
the slot of the existing delalloc extent to the new real extent before
inserting the shortened delalloc extent before or after it. This
avoids the overlapping extents while still allowing to update the
br_startblock field of the delalloc extent with the updated indirect
block reservation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Some were missed in the pass that converted the function return
values from int to bool. Update the remaining ones for consistency.
Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Move the error injection tag names into a libxfs header so that we can
share it between kernel and userspace.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Remove xfs_inode_log_format_t now that xfs_inode_log_format is
explicitly padded and therefore is a real on-disk structure. This
enables xfs/122 to check the size of the structure.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
[sandeen: same treatment in libxlog & logprint/ ] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
xfs: remove the inode log format from the inode log item
No need to keep the inode log format around all the time, we can
easily generate it at iop_format time.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Dave Chinner <david@fromorbit.com>
[sandeen: matching change in userspace xfs_trans.h] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Variable bit is being assigned a value that is never read, hence
the assignment is redundant and can be removed. Cleans up clang
warning:
fs/xfs/libxfs/xfs_rtbitmap.c:675:3: warning: Value stored to
'bit' is never read
Signed-off-by: Colin Ian King <colin.king@canonical.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
When we're done checking all the records/keys in a btree block, compute
the low and high key of the block and compare them to the associated key
in the parent btree block.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Abort an dir/attr btree operation if the attr btree has obvious problems
like loops back to the root or pointers don't point down the tree.
Found by fuzzing btree[0].before to zero in xfs/402, which livelocks on
the cycle in the attr btree.
Apply the same checks to xfs_da3_node_lookup_int.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This helper looks up the last extent the covers space before the passed
in block number. This is useful for truncate and similar operations that
operate backwards over the extent list. For xfs_bunmapi it also is
a slight optimization as we can return early if there are not extents
at or below the end of the to be truncated range.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
xfs_iread_extents is just a trivial wrapper, there is no good reason
to keep the two separate.
[darrick: minor fixups having left xfs_bmbt_validate_extent intact]
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Look at the return value of xfs_iext_get_extent instead of figuring out
the extent count first and looping up to it.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Rewrite xfs_bmap_insert_extents so that we don't rely on extent indices
except for iterating over them. Not being able to iterate to the previous
extent or finding the extent that stop_fsb is in are sufficient exit
conditions, and we don't need to do any extent count games given that:
a) we already flushed all delalloc extents past our start offset
before doing the operation
b) xfs_iext_count() includes delalloc extents anyway
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Rewrite xfs_bmap_collapse_extents so that we don't rely on extent indices
except for iterating over them. Not being able to iterate to the next
extent is a sufficient exit condition, and we don't need to do any extent
count games given that:
a) we already flushed all delalloc extents past our start offset
before doing the operation
b) xfs_iext_count() includes delalloc extents anyway
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This way the caller gets the proper updated extent returned in got.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Instead do the actual left and right shift work in the callers, and just
keep a helper to update the bmap and rmap btrees as well as the in-core
extent list.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Have a separate helper for insert vs collapse, as this prepares us for
simplifying the code in the next patches.
Also changed the done output argument to a bool intead of int for both
new functions.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
The define was always set to 1, which means looping until we reach is
was dead code from the start.
Also remove an initialization of next_fsb for the done case that doesn't
fit the new code flow - it was never checked by the caller in the done
case to start with.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
We can simply use the i_rdev field in the Linux inode and just convert
to and from the XFS dev_t when reading or logging/writing the inode.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
[sandeen: add i_rdev to our mocked up inode in userspace]
[djwong: use IRIX_DEV macros, assign in i_rdev in libxfs_ialloc] Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Remove the dead code dealing with the UUID fork format that was never
implemented in Linux (and neither in IRIX as far as I know).
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[sandeen: use __u8 not u8 for now, adapt xfs_logprint] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Instead of looping over all extents in some debug-only helper just
insert trace points into the loops that already exist in the calling
functions.
Also split the xfs_extlist trace point into one each for reading and
writing extents from disk.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
[sandeen: userspace-specific tracepoint macro updates] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
xfs_iext_update_extent already has basically all the information needed
to centralize the bmap pre/post tracing. We just need to pass inode +
bmap state instead of the inode fork pointer to get all trace annotations.
In addition to covering all the existing trace points this gives us
tracing coverage for the extent shifting operations for free.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Now that we use xfs_iext_insert this is already covered by the tracing
in that function.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
We already have all the information about the fork a=D1=95 well as additional
tracing information, so pass that to xfs_iext_remove().
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This creates the right initial bmap state from the passed in inode
fork enum.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Perform some quick sanity testing of the disk quota information.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Perform simple tests of the realtime bitmap and summary.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Scrub parent pointers, sort of. For directories, we can ride the
'..' entry up to the parent to confirm that there's at most one
dentry that points back to this directory.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create the infrastructure to scrub symbolic link data.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Scrub the hash tree, keys, and values in an extended attribute structure.
Refactor the attribute code to use the transaction if the caller supplied
one to avoid buffer deadocks.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Scrub the hash tree and all the entries in a directory.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Scrub an individual inode's block mappings to make sure they make sense.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Plumb in the pieces necessary to check the refcount btree. If rmap is
available, check the reference count by performing an interval query
against the rmapbt.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Check the reverse mapping records to make sure that the contents
make sense.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Check the records of the inode btrees to make sure that the values
make sense given the inode records themselves.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Check the extent records free space btrees to ensure that the values
look sane.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Add a forgotten check to the AGI verifier, then wire up the scrub
infrastructure to check the AGI contents.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Check the block references in the AGF and AGFL headers to make sure
they make sense.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Ensure that the geometry presented in the backup superblocks matches
the primary superblock so that repair can recover the filesystem if
that primary gets corrupted.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create a function that can check the shape of a btree -- each block
passes basic inspection and all the pointers look ok. In the next patch
we'll add the ability to check the actual keys and records stored within
the btree. Add some helper functions so that we report detailed scrub
errors in a uniform manner in dmesg. These are helper functions for
subsequent patches.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create a probe scrubber with id 0. This will be used by xfs_scrub to
probe the kernel's abilities to scrub (and repair) the metadata. We do
this by validating the ioctl inputs from userspace, preparing the
filesystem for a scrub (or a repair) operation, and immediately
returning to userspace. Userspace can use the returned errno and
structure state to decide (in broad terms) if scrub/repair are
supported by the running kernel.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create an ioctl that can be used to scrub internal filesystem metadata.
The new ioctl takes the metadata type, an (optional) AG number, an
(optional) inode number and generation, and a flags argument. This will
be used by the upcoming XFS online scrub tool.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create some helper functions to check that inode pointers point to
somewhere within the filesystem and not at the static AG metadata.
Move xfs_internal_inum and create a directory inode check function.
We will use these functions in scrub and elsewhere.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Refactor the btree block header checks to have an internal function that
returns the address of the failing check without logging errors. The
scrubber will call the internal function, while the external version
will maintain the current logging behavior.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Refactor the btree pointer checks so that we can call them from the
scrub code without logging errors to dmesg. Preserve the existing error
reporting for regular operations.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Create some helper functions to check that a block pointer points
within the filesystem (or AG) and doesn't point at static metadata.
We will use this for scrub.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
We only use xfs_bmbt_lookup_ge to look up the first bmap record in an
inode, so replace xfs_bmbt_lookup_ge with a special purpose helper that
is a bit more descriptive.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Now that we've massaged the callers into the right form we can always
pass the actual extent record instead of the individual fields.
As an additional benefit the btree cursor will now be prepoulated with
the correct extent state instead of having to fix it up later.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Now that we've massaged the callers into the right form we can always
pass the actual extent record instead of the individual fields.
With that xfs_bmbt_disk_set_allf can go away, and xfs_bmbt_disk_set_all
can be merged into the former implementation of xfs_bmbt_disk_set_allf.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use xfs_iext_get_extent to find, and xfs_iext_update_extent to update
entries in the in-core extent list. This isolates the function from
the detailed layout of the extent list, and generally makes the code
a lot more readable.
Also get rid of the oldext and newext variables as using the extent
records is a lot more descriptive.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Account for all changes to the delalloc reservation in da_new, and use a
single call xfs_mod_fdblocks to reserve/free blocks, including always
checking for an error.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use xfs_iext_get_extent to find, and xfs_iext_update_extent to update
entries in the in-core extent list. This isolates the function from
the detailed layout of the extent list, and generally makes the code
a lot more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use xfs_iext_update_extent to update entries in the in-core extent list.
This isolates the function from the detailed layout of the extent list,
and generally makes the code a lot more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use xfs_iext_get_extent to find, and xfs_iext_update_extent to update
entries in the in-core extent list. This isolates the function from
the detailed layout of the extent list, and generally makes the code
a lot more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use xfs_iext_update_extent to update entries in the in-core extent list.
This isolates the function from the detailed layout of the extent list,
and generally makes the code a lot more readable.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use the same defines as the other extent add and delete helpers, which
both improves code readability and trace point output.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Use the _FILLING values to match the usage in the xfs_bmap_add_extent_*
helpers. No change in behavior, just better naming in the code and
tracepoint output.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reported-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
And remove the delalloc code from xfs_bmap_del_extent, which gets renamed
to xfs_bmap_del_extent_real to fit the naming scheme used by the other
xfs_bmap_{add,del}_extent_* routines.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Rename the bno variable that's used as the end of the range in
__xfs_bunmapi to end, which better describes it. Additionally change
the start variable which takes the initial value of bno to be the
function parameter itself.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
The XFS_BTCUR_BPRV_WASDEL flag is supposed to indicate that we are
converting a delayed allocation to a real one, which isn't the case
in xfs_bunmapi. Setting it could theoretically lead to misaccounting
here, but it's unlikely that we ever hit it in practice.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This avoids exposure to details of the extent list implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
There was one spot in xfs_bmap_add_extent_unwritten_real that didn't use the
passed in new extent state but always converted to normal, leading to wrong
behavior when converting from normal to unwritten.
Only found by code inspection, it seems like this code path to move partial
extent from written to unwritten while merging it with the next extent is
rarely exercised.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Partial writes are performed when there is an error midway
while performing the I/O. Perform the write exactly once and
return the number of bytes written so far, until the error.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
[sandeen@sandeen.net: add O to short help] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This allows to make pwritev2() calls with RWF_NOWAIT,
which would fail in case the call blocks.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
[sandeen@sandeen.net: combine all optional args in short help] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
This results in default case which will display the error message
that command is not found. This is required for xfs_io builds which
did not support pwritev and an attempt is made to use -V.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com>
[sandeen@sandeen.net: restrict default: printf to ifdef'd options] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:30 +0000 (11:35 -0600)]
xfs_repair: always release btree buffers when refcountbt checks fail
During phase 4 of repair, we compare the refcount records we've computed
against what's in the refcount btree. If the btree is corrupt the
libxfs refcount calls will fail, but on our way out of the function we
don't make sure to clear all the btree cursor's buffer pointers. This
leads to a repair deadlock in xfs/372 when fuzzing ptrs[1] to -1U
because we forgot to release a buffer.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:29 +0000 (11:35 -0600)]
misc: fix ubsan warnings
Fix all the things ubsan warned about.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:28 +0000 (11:35 -0600)]
misc: enable thread sanitizer if the builder wants it
Enable the gcc/clang thread data corruption sanitizer if the builder
requests it and it's available.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:27 +0000 (11:35 -0600)]
misc: enable gcc/clang address sanitizer if the builder wants it
Enable AddressSanitizer to look for memory usage errors if the builder
asks for it and it's available.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:26 +0000 (11:35 -0600)]
misc: enable ubsan if the builder wants it
Enable the undefined behavior sanitizer (ubsan) if the builder requests
it and it's available.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:26 +0000 (11:35 -0600)]
xfs_repair: fix cowextsize field checking and repairing
Make sure that we never leave the filesystem with a zero cowextsize hint
while the cowextsize inode flag is set.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:25 +0000 (11:35 -0600)]
xfs_repair: clear DAX flag from non-file inodes
The DAX flag should only be set for files and directories.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:24 +0000 (11:35 -0600)]
xfs_repair: fix bag memory overwrite problems
There's an off by one error in the bag_remove code such that we end up
copying memory from beyond the end of the array into the array. Not a
serious problem since we have counters to prevent us from reading that
garbage, but AddressSanitizer complained so let's fix it.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:23 +0000 (11:35 -0600)]
xfs_io: add new error injection knobs to inject command
Bring xfs_io's inject command up to date with the newest knobs.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:23 +0000 (11:35 -0600)]
xfsprogs: explicitly cast troublesome types to match printf format specifiers
Certain system-defined types (__u64, __s64, __nlink_t, __ino64_t,
__off64_t, __blkcnt64_t) don't have a consistent definition across
different architectures, so wherever we use a printf format specifier on
such a variable, we have to typecast the variable or else the compiler
will complain.
IOWs this fixes build warnings on ppc64le.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Darrick J. Wong [Thu, 9 Nov 2017 17:35:22 +0000 (11:35 -0600)]
db: increase metadump's default overly long extent discard threshold
Back in 88b8e1d6d7 ("Make xfs_metadump more robust against bad data"),
metadump grew the ability to ignore a directory extent if it was longer
than 20 blocks. Presumably this was to protect metadump from dumping
absurdly long extents resulting from bmbt corruption, but it's certainly
possible to create a directory with an extent longer than 20 blocks.
Hilariously, the discards happen with no warning unless the caller
explicitly set -w.
This was raised to 1000 blocks in 7431d134fe8 ("Increase default maximum
extent size for xfs_metadump when copying..."), but it's still possible
to create a directory with an extent longer than 1000 blocks.
Increase the threshold to MAXEXTLEN blocks because it's totally valid
for the filesystem to create extents up to that length.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
build: define _DEFAULT_SOURCE with _BSD_SOURCE for preadv
_BSD_SOURCE has been deprecated and replaced with _DEFAULT_SOURCE
for feature test macros. We use this for preadv in the configure
script, but the latest preadv(2) manpage says:
preadv(), pwritev():
since glibc 2.19:
_DEFAULT_SOURCE
Glibc 2.19 and earlier:
_BSD_SOURCE
and the latest feature_test_macros(7) manpage says:
Since glibc 2.20, this macro is deprecated. It now has the same
effect as defining _DEFAULT_SOURCE, but generates a compile-time
warning (unless _DEFAULT_SOURCE is also defined). Use
_DEFAULT_SOURCE instead. To allow code that requires
_BSD_SOURCE in glibc 2.19 and earlier and _DEFAULT_SOURCE in
glibc 2.20 and later to compile without warnings, define both
_BSD_SOURCE and _DEFAULT_SOURCE.
So add both to our configure test. They were completely missing
from the actual preadv usage in xfs_io, so add them there too.
Signed-off-by: Luis R. Rodriguez <mcgrof@kernel.org> Reviewed-by: Eric Sandeen <sandeen@redhat.com>
[sandeen@redhat.com: clarify changelog, add defines to code use too] Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Eric Sandeen [Thu, 9 Nov 2017 17:35:19 +0000 (11:35 -0600)]
xfs_metadump: zap stale data in DIR2_LEAF1 dirs
xfs_metadump attempts to zero out unused regions of metadata
blocks to prevent data leaks when sharing metadata images.
However, Stefan Ring reported a significant number of leaked
strings when dumping his 1T filesystem. Based on a reduced
metadata set, I was able to identify "leaf" directories
(with XFS_DIR2_LEAF1_MAGIC magic) as the primary culprit;
the region between the end of the entries array and the start
of the bests array was not getting zeroed out. This patch
seems to remedy that problem.
Reported-by: Stefan Ring <stefanrin@gmail.com> Signed-off-by: Eric Sandeen <sandeen@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Jan Tulak [Thu, 9 Nov 2017 17:35:18 +0000 (11:35 -0600)]
metadump: update manpage for dirty log/obfuscation issue
This is something that should be documented, as it is not obvious to
everyone.
Signed-off-by: Jan Tulak <jtulak@redhat.com>
[sandeen@redhat.com: edit for clarity]
[darrick.wong@oracle.com: edit for more clarity] Reviewed-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: Eric Sandeen <sandeen@sandeen.net>