Make things more consistent and ensure that we're using u64 bitfields -
key types and btree ids are already around 32 bits.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
the order in which btree_gc walks keys have changed, so we no longer
have the sort of issues with online fsck this assertion was warning
about.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Rewrite fsck/gc for the new accounting scheme.
This adds a second set of in-memory accounting counters for gc to use;
like with other parts of gc we run all trigger in TRIGGER_GC mode, then
compare what we calculated to existing in-memory accounting at the end.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Main part of the disk accounting rewrite.
This is a wholesale rewrite of the existing disk space accounting, which
relies on percepu counters that are sharded by journal buffer, and
rolled up and added to each journal write.
With the new scheme, every set of counters is a distinct key in the
accounting btree; this fixes scaling limitations of the old scheme,
where counters took up space in each journal entry and required multiple
percpu counters.
Now, in memory accounting requires a single set of percpu counters - not
multiple for each in flight journal buffer - and in the future we'll
probably also have counters that don't use in memory percpu counters,
they're not strictly required.
An accounting update is now a normal btree update, using the btree write
buffer path. At transaction commit time, we apply accounting updates to
the in memory counters, which are percpu counters indexed in an
eytzinger tree by the accounting key.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Until accounting keys hit the btree, they are deltas, not new versions
of the existing key; this means we have to teach journal replay to
accumulate them.
Additionally, the journal doesn't track precisely which entries have
been flushed to the btree; it only tracks a range of entries that may
possibly still need to be flushed.
That means we need to compare accounting keys against the version in the
btree and only flush updates that are newer.
There's another wrinkle with the write buffer: if the write buffer
starts flushing accounting keys before journal replay has finished
flushing accounting keys, journal replay will see the version number
from the new updates and updates from the journal will be lost.
To avoid this, journal replay has to flush accounting keys first, and
we'll be adding a flag so that write buffer flush knows to hold
accounting keys until then.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Use try_cmpxchg() family of functions instead of
cmpxchg (*ptr, old, new) == old. x86 CMPXCHG instruction returns
success in ZF flag, so this change saves a compare after cmpxchg
(and related move instruction in front of cmpxchg).
Also, try_cmpxchg() implicitly assigns old *ptr value to "old" when
cmpxchg fails. There is no need to re-read the value in the loop.
No functional change intended.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Brian Foster <bfoster@redhat.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We're about to start using bch_validate_flags for superblock section
validation - it's no longer bkey specific.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Combine iter/update/trigger/str_hash flags into a single enum, and
x-macroize them for a to_text() function later.
These flags are all for a specific iter/key/update context, so it makes
sense to group them together - iter/update/trigger flags were already
given distinct bits, this cleans up and unifies that handling.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
btree_key_can_insert_cached() should be checking the watermark -
BCH_TRANS_COMMIT_journal_replay really means nonblocking mode when
watermark < reclaim, it was being used incorrectly.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
The btree paths array is now dynamically resizable - and as well the
btree_insert_entries array, as it needs to be the same size.
The merge path (and interior update path) allocates new btree paths,
thus can trigger a resize; thus we need to not retain direct pointers
after invoking merge; similarly when running btree node triggers.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This adds a new watermark, higher priority than BCH_WATERMARK_reclaim,
for interior btree updates. We've seen a deadlock where journal replay
triggers a ton of btree node merges, and these use up all available open
buckets and then interior updates get stuck.
One cause of this is that we're currently lacking btree node merging on
write buffer btrees - that needs to be fixed as well.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Previously, we assumed that keys were consistent with the snapshots
btree - but that's not correct as fsck may not have been run or may not
be complete.
This adds checks and error handling when using the in-memory snapshots
table (that mirrors the snapshots btree).
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
bcachefs btree nodes are big - typically 256k - and btree roots are
pinned in memory. As we're now up to 18 btrees, we now have significant
memory overhead in mostly empty btree roots.
And in the future we're going to start enforcing that certain btree node
boundaries exist, to solve lock contention issues - analagous to XFS's
AGIs.
Thus, we need to start allocating smaller btree node buffers when we
can. This patch changes code that refers to the filesystem constant
c->opts.btree_node_size to refer to the btree node buffer size -
btree_buf_bytes() - where appropriate.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
The disk space accounting rewrite is splitting out accounting for each
replicas set - those are moving to btree keys, instead of percpu
counters.
This breaks bch2_trans_fs_usage_apply() up, splitting out the part we
will still need.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Prep work for disk space accounting rewrite: we're going to want to use
a single callback for both of our current triggers, so we need to change
them to have the same type signature first.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Prep work for disk space accounting rewrite: we're going to want to use
a single callback for both of our current triggers, so we need to change
them to have the same type signature first.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Since the btree_paths array is now about to become growable, we have to
be careful not to refer to paths by pointer across contexts where they
may be reallocated.
This fixes the remaining btree_interior_update() paths - split and
merge.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Previosuly, the transaction commit path would have to add keys to the
btree write buffer as a separate operation, requiring additional global
synchronization.
This patch introduces a new journal entry type, which indicates that the
keys need to be copied into the btree write buffer prior to being
written out. We switch the journal entry type back to
JSET_ENTRY_btree_keys prior to write, so this is not an on disk format
change.
Flushing the btree write buffer may require pulling keys out of journal
entries yet to be written, and quiescing outstanding journal
reservations; we previously added journal->buf_lock for synchronization
with the journal write path.
We also can't put strict bounds on the number of keys in the journal
destined for the write buffer, which means we might overflow the size of
the preallocated buffer and have to reallocate - this introduces a
potentially fatal memory allocation failure. This is something we'll
have to watch for, if it becomes an issue in practice we can do
additional mitigation.
The transaction commit path no longer has to explicitly check if the
write buffer is full and wait on flushing; this is another performance
optimization. Instead, when the btree write buffer is close to full we
change the journal watermark, so that only reservations for journal
reclaim are allowed.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Instead of using a darray, we now allocate journal entries for the
transaction commit path with our normal bump allocator - with an inlined
fastpath, and using btree_transaction_stats to remember how much to
initially allocate so as to avoid transaction restarts.
This is prep work for converting write buffer updates to use this
mechanism.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Minor refactoring - improved naming, and move the responsibility for
flush_lock to the caller instead of having it be shared.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
__bch2_btree_write_buffer_flush() now assumes a write ref is already
held (as called by the transaction commit path); and the wrappers
bch2_write_buffer_flush() and flush_sync() take an explicit write ref.
This means internally the write buffer code can always use
BTREE_INSERT_NOCHECK_RW, instead of in the previous code passing flags
around and hoping the NOCHECK_RW flag was always carried around
correctly.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Now we can print out filesystem flags in sysfs, useful for debugging
various "what's my filesystem doing" issues.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
With the previous patch that reworks BTREE_INSERT_JOURNAL_REPLAY, we can
now switch the btree write buffer to use it for flushing.
This has the advantage that transaction commits don't need to take a
journal reservation at all.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This slightly changes how trans->journal_res works, in preparation for
changing the btree write buffer flush path to use it.
Now, BTREE_INSERT_JOURNAL_REPLAY means "don't take a journal
reservation; trans->journal_res.seq already refers to the journal
sequence number to pin".
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
The upcoming btree write buffer rework is going to use the journal
itself as the first stage of the write buffer; this is a cleanup to make
sure k->needs_whiteout is initialized before keys hit the journal.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This deletes the complicated and somewhat expensive journal
pre-reservation machinery in favor of just using journal watermarks:
when the journal is more than half full, we run journal reclaim more
aggressively, and when the journal is more than 3/4s full we only allow
journal reclaim to get new journal reservations.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We really don't want to be invoking memory reclaim with btree locks
held: even aside from (solvable, but tricky) recursion issues, it can
cause painful to diagnose performance edge cases.
This fixes a recently reported issue in btree_key_can_insert_cached().
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Reported-by: Mateusz Guzik <mjguzik@gmail.com>
Fixes: https://lore.kernel.org/linux-bcachefs/CAGudoHEsb_hGRMeWeXh+UF6po0qQuuq_NKSEo+s1sEb6bDLjpA@mail.gmail.com/T/
As prep work for the next patch to fix a key cache reclaim issue, we
need to start tracking whether we're currently holding write locks - so
that we can release and retake the before calling into memory reclaim.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
rebalance_work entries may refer to entries in the extents btree, which
is a snapshots btree, or they may also refer to entries in the reflink
btree, which is not.
Hence rebalance_work keys may use the snapshot field but it's not
required to be nonzero - add a new btree flag to reflect this.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
gcc 10 seems to complain about array bounds in situations where gcc 11
does not - curious.
This unfortunately requires adding some casts for now; we may
investigate getting rid of our __u64 _data[] VLA in a future patch so
that our start[0] members can be VLAs.
Reported-by: John Stoffel <john@stoffel.org>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
We should only be downgrading locks on success - otherwise, our
transaction restarts won't be getting the correct locks and we'll
livelock.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Upcoming rebalance_work btree will require extent triggers to be
BTREE_TRIGGER_WANTS_OLD_AND_NEW - so to reduce potential confusion,
let's just make all triggers BTREE_TRIGGER_WANTS_OLD_AND_NEW.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
