Kent Overstreet [Sat, 3 Apr 2021 20:24:13 +0000 (16:24 -0400)]
bcachefs: Be more careful about JOURNAL_RES_GET_RESERVED
JOURNAL_RES_GET_RESERVED should only be used for updatse that need to be
done to free up space in the journal. In particular, when we're flushing
keys from the key cache, if we're flushing them out of order we
shouldn't be using it, since we're using up our remaining space in the
journal without dropping a pin that will let us make forward progress.
With this patch, BTREE_INSERT_JOURNAL_RECLAIM without
BTREE_INSERT_JOURNAL_RESERVED may return -EAGAIN - we can't wait on
journal reclaim if we're already in journal reclaim.
This means we need to propagate these errors up to journal reclaim,
indicating that flushing a journal pin should be retried in the future.
This is prep work for a patch to change the way journal reclaim works,
to split out flushing key cache keys because the btree key cache is too
dirty from journal reclaim because we need space in the journal.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 3 Apr 2021 23:27:05 +0000 (19:27 -0400)]
bcachefs: Fix journal deadlock
After we get a journal reservation, we need to use it - if we erorr out
of a transaction commit, we'll be eating into space in the journal and
if our transaction needs to make forward progress in order to reclaim
space in the journal, we'll deadlock.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 3 Apr 2021 01:29:05 +0000 (21:29 -0400)]
bcachefs: Increase commality between BTREE_ITER_NODES and BTREE_ITER_KEYS
Eventually BTREE_ITER_NODES should be going away. This patch is to fix a
transaction iterator overflow in the btree node merge path because
BTREE_ITER_NODES iterators couldn't be reused.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 29 Mar 2021 05:13:31 +0000 (01:13 -0400)]
bcachefs: Move btree node merging to before transaction commit
Currently, BTREE_INSERT_NOUNLOCK makes it hard to ensure btree node
merging happens reliably - since btree node merging happens after
transaction commit, we can't drop btree locks and block when starting
the btree update.
This patch moves it to before transaction commit - and failure to do a
merge that we wanted to do just restarts the transaction.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 29 Mar 2021 05:13:31 +0000 (01:13 -0400)]
bcachefs: Kill bch2_btree_node_get_sibling()
This patch reworks the btree node merge path to use a second btree
iterator to get the sibling node - which means
bch2_btree_iter_get_sibling() can be deleted. Also, it uses
bch2_btree_iter_traverse_all() if necessary - which means it should be
more reliable. We don't currently even try to make it work when
trans->nounlock is set - after a BTREE_INSERT_NOUNLOCK transaction
commit, hopefully this will be a worthwhile tradeoff.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Wed, 31 Mar 2021 19:39:16 +0000 (15:39 -0400)]
bcachefs: Change where merging of interior btree nodes is trigger from
Previously, we were doing btree node merging from
bch2_btree_insert_node() - but this is called from the split path, when
we're in the middle of creating new nodes and deleting new nodes and the
iterators are in a weird state.
Also, this means we're starting a new btree_update while in the middle
of an existing one, and that's asking for deadlocks.
Much simpler and saner to trigger btree node merging _after_ the whole
btree node split path is finished.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Wed, 31 Mar 2021 19:21:37 +0000 (15:21 -0400)]
bcachefs: Improve bch2_btree_update_start()
bch2_btree_update_start() is now responsible for taking gc_lock and
upgrading the iterator to lock parent nodes - greatly simplifying error
handling and all of the callers.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 29 Mar 2021 04:19:05 +0000 (00:19 -0400)]
bcachefs: Add repair code for out of order keys in a btree node.
This just drops the offending key - in the bug report where this was
seen, it was clearly a single bit memory error, and fsck will fix the
missing key.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 29 Mar 2021 00:57:59 +0000 (20:57 -0400)]
bcachefs: Have journal reclaim thread flush more aggressively
This adds a new watermark for the journal reclaim when flushing btree
key cache entries - it should try and stay ahead of where foreground
threads doing transaction commits will enter direct journal reclaim.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 16 Mar 2021 22:08:10 +0000 (18:08 -0400)]
bcachefs: Don't use bch2_inode_find_by_inum() in move.c
Since move.c isn't aware of what subvolume we're in, we can't use the
standard inode lookup code - fortunately, we're just using it for
reading IO options.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Mar 2021 23:18:30 +0000 (19:18 -0400)]
bcachefs: Change inode allocation code for snapshots
For snapshots, when we allocate a new inode we want to allocate an inode
number that isn't in use in any other subvolume. We won't be able to use
ITER_SLOTS for this, inode allocation needs to change to use
BTREE_ITER_ALL_SNAPSHOTS.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 2 Mar 2021 23:35:30 +0000 (18:35 -0500)]
bcachefs: Inode backpointers
This patch adds two new inode fields, bi_dir and bi_dir_offset, that
point back to the inode's dirent.
Since we're only adding fields for a single backpointer, files that have
been hardlinked won't necessarily have valid backpointers: we also add a
new inode flag, BCH_INODE_BACKPTR_UNTRUSTED, that's set if an inode has
ever had multiple links to it. That's ok, because we only really need
this functionality for directories, which can never have multiple
hardlinks - when we add subvolumes, we'll need a way to enemurate and
print subvolumes, and this will let us reconstruct a path to a subvolume
root given a subvolume root inode.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Wed, 24 Mar 2021 22:02:16 +0000 (18:02 -0400)]
bcachefs: Start using bpos.snapshot field
This patch starts treating the bpos.snapshot field like part of the key
in the btree code:
* bpos_successor() and bpos_predecessor() now include the snapshot field
* Keys in btrees that will be using snapshots (extents, inodes, dirents
and xattrs) now always have their snapshot field set to U32_MAX
The btree iterator code gets a new flag, BTREE_ITER_ALL_SNAPSHOTS, that
determines whether we're iterating over keys in all snapshots or not -
internally, this controlls whether bkey_(successor|predecessor)
increment/decrement the snapshot field, or only the higher bits of the
key.
We add a new member to struct btree_iter, iter->snapshot: when
BTREE_ITER_ALL_SNAPSHOTS is not set, iter->pos.snapshot should always
equal iter->snapshot, which will be 0 for btrees that don't use
snapshots, and alsways U32_MAX for btrees that will use snapshots
(until we enable snapshot creation).
This patch also introduces a new metadata version number, and compat
code for reading from/writing to older versions - this isn't a forced
upgrade (yet).
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Thu, 4 Mar 2021 21:20:16 +0000 (16:20 -0500)]
bcachefs: Split out bpos_cmp() and bkey_cmp()
With snapshots, we're going to need to differentiate between comparisons
that should and shouldn't include the snapshot field. bpos_cmp is now
the comparison function that does include the snapshot field, used by
core btree code.
Upper level filesystem code generally does _not_ want to compare against
the snapshot field - that code wants keys to compare as equal even when
one of them is in an ancestor snapshot.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Thu, 25 Mar 2021 03:37:33 +0000 (23:37 -0400)]
bcachefs: btree key cache locking improvements
The btree key cache mutex was becoming a significant bottleneck - it was
mainly used to protect the lists of dirty, clean and freed cached keys.
This patch eliminates the dirty and clean lists - instead, when we need
to scan for keys to drop from the cache we iterate over the rhashtable,
and thus we're able to remove most uses of that lock.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sun, 28 Mar 2021 00:58:57 +0000 (20:58 -0400)]
bcachefs: Fix for bch2_trans_commit() unlocking when it's not supposed to
When we pass BTREE_INSERT_NOUNLOCK bch2_trans_commit isn't supposed to
unlock after a successful commit, but it was calling
bch2_trans_cond_resched() - oops.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 27 Mar 2021 00:08:56 +0000 (20:08 -0400)]
bcachefs: Generate better bkey formats when splitting nodes
On btree node split, we weren't ensuring the min_key of the new larger
node packs in the new format for this node. This triggers some painful
slowpaths in the bset.c aux search tree code - this patch fixes that by
calculating a new format for the new node with the new min_key.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Thu, 25 Mar 2021 00:22:51 +0000 (20:22 -0400)]
bcachefs: Drop bkey noops
Bkey noops were introduced to deal with trimming inline data extents in
place in the btree: if the u64s field of a bkey was 0, that u64 was a
noop and we'd start looking for the next bkey immediately after it.
But extent handling has been lifted above the btree - we no longer
modify existing extents in place in the btree, and the compatibilty code
for old style extent btree nodes is gone, so we can completely drop this
code.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Wed, 24 Mar 2021 01:22:50 +0000 (21:22 -0400)]
bcachefs: Split btree_iter_traverse and bch2_btree_iter_traverse()
External (to the btree iterator code) users of bch2_btree_iter_traverse
expect that on success the iterator will be pointed at iter->pos and
have that position locked - but since we split iter->pos and
iter->real_pos, that means it has to update iter->real_pos if necessary.
Internal users don't expect it to modify iter->real_pos, so we need two
separate functions.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sun, 21 Mar 2021 22:09:02 +0000 (18:09 -0400)]
bcachefs: Have btree_iter_next_node() use btree_iter_set_search_pos()
btree node iterators need to obey the regular btree node invarionts
w.r.t. iter->real_pos; once they do, bch2_btree_iter_traverse will have
less that it needs to check.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sun, 21 Mar 2021 20:55:25 +0000 (16:55 -0400)]
bcachefs: Replace bch2_btree_iter_next() calls with bch2_btree_iter_advance
The way btree iterators work internally has been changing, particularly
with the iter->real_pos changes, and bch2_btree_iter_next() is no longer
hyper optimized - it's just advance followed by peek, so it's more
efficient to just call advance where we're not using the return value of
bch2_btree_iter_next().
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 20 Mar 2021 03:19:05 +0000 (23:19 -0400)]
bcachefs: Kill bkey ops->debugcheck method
This code used to be used for running some assertions on alloc info at
runtime, but it long predates fsck and hasn't been good for much in
ages - we can delete it now.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 20 Mar 2021 02:34:54 +0000 (22:34 -0400)]
bcachefs: Fsck code refactoring
Change fsck code to always put btree iterators - also, make some flow
control improvements to deal with lock restarts better, and refactor
check_extents() to not walk extents twice for counting/checking
i_sectors.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Mar 2021 21:26:19 +0000 (17:26 -0400)]
bcachefs: Kill reflink option
An option was added to control whether reflink support was on or off
because for a long time, reflink + inline data extent support was
missing - but that's since been fixed, so we can drop the option now.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 15 Mar 2021 01:30:08 +0000 (21:30 -0400)]
bcachefs: Fix read retry path for indirect extents
In the read path, for retry of indirect extents to work we need to
differentiate between the location in the btree the read was for, vs.
the location where we found the data. This patch adds that plumbing to
bch_read_bio.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Mon, 8 Mar 2021 22:09:13 +0000 (17:09 -0500)]
bcachefs: Fix locking in bch2_btree_iter_traverse_cached()
bch2_btree_iter_traverse() is supposed to ensure we have the correct
type of lock - it was downgrading if necessary, but if we entered with a
read lock it wasn't upgrading to an intent lock, oops.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sun, 21 Feb 2021 01:51:57 +0000 (20:51 -0500)]
bcachefs: Improve handling of extents in bch2_trans_update()
The transaction update/commit path cares about whether it's inserting
extents or regular keys; extents require extra passes (handling of
overlapping extents) but sometimes we want to skip all that. This
clarifies things by adding a new member to btree_insert_entry specifying
whether the key being inserted is an extent, instead of overloading
BTREE_ITER_IS_EXTENTS.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 20 Feb 2021 04:41:40 +0000 (23:41 -0500)]
bcachefs: KEY_TYPE_discard is no longer used
KEY_TYPE_discard used to be used for extent whiteouts, but when handling
over overlapping extents was lifted above the core btree code it became
unused. This patch updates various code to reflect that.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 20 Feb 2021 05:00:23 +0000 (00:00 -0500)]
bcachefs: Kill support for !BTREE_NODE_NEW_EXTENT_OVERWRITE()
bcachefs has been aggressively migrating filesystems and btree nodes to
the new format for quite some time - this shouldn't affect anyone
anymore, and lets us delete a _lot_ of code. Also, it frees up
KEY_TYPE_discard for a new whiteout key type for snapshots.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 28 Dec 2021 03:11:54 +0000 (22:11 -0500)]
bcachefs: Fix bch2_btree_cache_scan()
It was counting nodes on the freed list that it skips - because we want
to leave a few so that btree splits don't touch the allocator - as nodes
that it touched, meaning that if it was called with <= 3 nodes to
reclaim, and those nodes were on the freed list, it would never do any
work.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sun, 18 Apr 2021 22:01:49 +0000 (18:01 -0400)]
bcachefs: Fix for copygc getting stuck waiting for reserve to be filled
This fixes a regression from the patch
bcachefs: Fix copygc dying on startup
In general only the allocator thread itself should be updating
ca->allocator_state, the thread waking up the allocator setting it is an
ugly hack only needed to avoid racing with the copygc threads when we're
first starting up.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Tue, 13 Apr 2021 13:49:23 +0000 (09:49 -0400)]
bcachefs: Fix copygc threshold
Awhile back the meaning of is_available_bucket() and thus also
bch_dev_usage->buckets_unavailable changed to include buckets that are
owned by the allocator - this was so that the stat could be persisted
like other allocation information, and wouldn't have to be regenerated
by walking each bucket at mount time.
This broke copygc, which needs to consider buckets that are reclaimable
and haven't yet been grabbed by the allocator thread and moved onta
freelist. This patch fixes that by adding dev_buckets_reclaimable() for
copygc and the allocator thread, and cleans up some of the callers a bit.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Fri, 16 Apr 2021 22:59:54 +0000 (18:59 -0400)]
bcachefs: Don't drop ptrs to btree nodes
If a ptr gen doesn't match the bucket gen, the bucket likely doesn't
contain the data we want - but it's still possible the data we want
might have been overwritten, and for btree node pointers we can verify
whether or not the node is the one we wanted with the node's sequence
number, so it's better to keep the pointer and try reading from it.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Fri, 16 Apr 2021 20:54:11 +0000 (16:54 -0400)]
bcachefs: Bring back metadata only gc
This is useful for the filesystem dump debugging tool - when we're
hitting bugs we want to skip as much of the recovery process as
possible, and the dump tool only needs to know where metadata lives.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Kent Overstreet [Sat, 3 Apr 2021 03:41:10 +0000 (23:41 -0400)]
bcachefs: Don't fail mounts due to devices that are marked as failed
If a given set of replicas is entirely on failed devices, don't fail the
mount: we will still fail the mount if we have some copies on non failed
devices.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
projects / mirror_ubuntu-kernels.git / log