This code needs quite a bit of work: we don't want to be walking all
metadata in the filesystem, we should just be walking backpointers, and
it should be switched to a data ioctl that can report progress via a
file descriptor, not the system console.
But that'll take more work - before we can safely walk only backpointers
we need to change device add to not reuse device indexes, since with
that change accounting being wrong introduces the possibility of
removing a device that still has pointers.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
the backpointers code has progress indicators; these aren't great, since
they print to the dmesg console and we much prefer to have progress
indicators reporting to a specific userspace program so they're not
spamming the system console.
But not all codepaths that need progress indicators support that yet,
and we don't want users to think "this is hung".
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
we're starting to use error messages with paths in fsck_errors(), where
we do not want nested transaction restart handling, so let's prepare for
that.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We want all error messages converted to print paths, not just inode
numbers - users want this information, and it speeds up debugging too.
Auditing and converting all error messages is going to be a big project,
so for the moment we're just doing this incrementally.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Iterating over backpointers on a specific device is potentially much
cheaper than walking all filesystem data.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Reorganize counters a bit, grouping related counters together.
New counters:
- io_read_inline
- io_read_hole
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
When ancestor is less than IS_ANCESTOR_BITMAP, we would get an incorrect
result.
Signed-off-by: Alan Huang <mmpgouride@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Add a new data op to walk all data and metadata in a filesystem,
checking if it can be read successfully, and on error repairing from
another copy if possible.
- New helper: bch2_dev_idx_is_online(), so that we can bail out and
report to userspace when we're unable to scrub because the device is
offline
- data_update_opts, which controls the data move path, now understands
scrub: data is only read, not written. The read path is responsible
for rewriting on read error, as with other reads.
- scrub_pred skips data extents that don't have checksums
- bch_ioctl_data has a new scrub member, which has a data_types field
for data types to check - i.e. all data types, or only metadata.
- Add new entries to bch_move_stats so that we can report numbers for
corrected and uncorrected errors
- Add a new enum to bch_ioctl_data_event for explicitly reporting
completion and return code (i.e. device offline)
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Add a function for scrubbing btree nodes - reading them in, and kicking
off a rewrite if there's an error.
The btree_node_read_done() checks have to be duplicated because we're
not using a pointer to a struct btree - the btree node might already be
in cache, and we need to check a specific replica, which might not be
the one we previously read from.
This will be used in the next patch implementing high-level scrub.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Add a new helper for rewriting a btree node given a just the key, not a
pointer to the node itself.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We may not need to pull in a btree node when walking backpointers -
don't do so unnecessarily when using backpointer_get_key().
It'll still fall back to backpointer_get_node() in a few situations,
including btree roots (where an iterator can't point at just the key),
and races due to the interior update path not having deleted a
backpointer to an old node yet.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Rework the read path so that BCH_READ_NODECODE reads now also self-heal
after a read error and a successful retry - prerequisite for scrub.
- __bch2_read_endio() now handles a read that's both BCH_READ_NODECODE
and a bounce.
Normally, we don't want a BCH_READ_NODECODE read to ever allocate a
split bch_read_bio: we want to maintain the relationship between the
bch_read_bio and the data_update it's embedded in.
But correcting read errors requires allocating a split/bounce rbio
that's embedded in a promote_op. We do still have a 1-1 relationship,
i.e. we only allocate a single split/bounce if it's a
BCH_READ_NODECODE, so things hopefully don't get too crazy.
- __bch2_read_extent() now is allowed to allocate the promote_op for
rewriting after a failed read, even if it's BCH_READ_NODECODE.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
we don't want to block completion of the read - starting a promote calls
into the write path, which will block.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
If a data update doesn't want to block on allocations (promotes, self
healing on read error) - check if the allocation would fail before
kicking off the data update and calling into the write path.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Initialize the write op first, so that in the next patch we can check if
the allocator would block (for BCH_WRITE_alloc_nowait ops) and bail out
before taking nocow locks/dev refs.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
If a drive is failing and we're moving data off of it, we can't
necessairly depend on capacity/disk reservation calculations to avoid
deadlocking/blocking on the allocator.
And, we don't want to queue up infinite self healing moves anyways.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Promotes, like most other internal moves, should only go to the
specified target and not fall back to allocating from the full
filesystem.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Now that data_update embeds bch_read_bio, BCH_READ_NODECODE means that
the read is embedded in a a data_update - and we can check in the retry
path if the extent has changed and bail out.
This likely fixes some subtle bugs with read errors and data moves.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Add a new helper for bch2_moving_ctxt_to_text(), which may be used to
debug if moving_ios are getting stuck.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Add an ioctl for querying counters, the same ones provided in
/sys/fs/bcachefs/<uuid>/counters/, but more suitable for a 'bcachefs
top' command.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Persistent counters, like recovery passes, include a stable enum in
their definition - but this was never correctly plumbed.
This allows us to add new counters and properly organize them with a
non-stable "presentation order", which can also be used in userspace by
the new 'bcachefs fs top' tool.
Fortunatel, since we haven't yet added any new counters where
presentation order ID doesn't match stable ID, this won't cause any
reordering issues.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We've got per-writepoint statistics to see how well the writepoint index
update threads are pipelining; this separates running vs. runnable so we
can see at a glance if they're blocking.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This makes 'bcachefs fs top' more useful; we can now see at a glance
whether the IO to the device is being done for user reads/writes, or
copygc/rebalance.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Early version of 'bcachefs_metadata_version_cached_backpointers' was
creating backpointers for stale cached pointers - whoops. Now we have to
repair those.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Factor out get_iter_to_node() and use it for
btree_node_rewrite_get_iter(), to be used for fixing btree node write
error behaviour.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
bcachefs removed most PAGE_SIZE references long ago, so this is easy;
only readpage_bio_extend() has to be tweaked to respect the minimum
order.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
bare 64 bit divides not allowed, whoops
arm-linux-gnueabi-ld: drivers/char/random.o: in function `__get_random_u64_below':
drivers/char/random.c:602:(.text+0xc70): undefined reference to `__aeabi_uldivmod'
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We just had a report of the assert for "btree in write buffer for
non-write buffer btree" popping during the 6.14 upgrade.
- 150TB filesystem, after a reboot the upgrade was able to continue from
where it left off, so no major damage.
But with 6.14 about to come out we want to get this tracked down asap,
and need more data if other users hit this.
Convert the BUG_ON() to an emergency read-only, and print out btree, the
key itself, and stack trace from the original write buffer update (which
did not have this check before).
Reported-by: Stijn Tintel <stijn@linux-ipv6.be>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
get_random_u32_below() has a better algorithm than bch2_rand_range(),
it just didn't exist at the time.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
