To implement snapshots, we need every filesystem btree operation (every
btree operation without a subvolume) to start by looking up the
subvolume and getting the current snapshot ID, with
bch2_subvolume_get_snapshot() - then, that snapshot ID is used for doing
btree lookups in BTREE_ITER_FILTER_SNAPSHOTS mode.
This patch adds those bch2_subvolume_get_snapshot() calls, and also
switches to passing around a subvol_inum instead of just an inode
number.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
For snapshots, we need to implement btree lookups that return the first
key that's an ancestor of the snapshot ID the lookup is being done in -
and filter out keys in unrelated snapshots. This patch adds the btree
iterator flag BTREE_ITER_FILTER_SNAPSHOTS which does that filtering.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Filesystem operations generally operate within a subvolume: at the start
of every btree transaction we'll be looking up (and locking) the
subvolume to get the current snapshot ID, which we then use for our
other btree lookups in BTREE_ITER_FILTER_SNAPSHOTS mode.
But inodes don't record what subvolume they're in - they can't, because
if they did we'd have to update every single inode within a subvolume
when taking a snapshot in order to keep that field up to date. So it
needs to be tracked in memory, based on how we got to that inode.
Hence this patch adds a subvolume field to ei_inode_info, and switches
to iget5() so we can index by it in the inode hash table.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
On existing filesystems, we have a single global lost+found. Introducing
subvolumes means we need to introduce per subvolume lost+found
directories, because inodes are added to lost+found by their inode
number, and inode numbers are now only unique within a subvolume.
This patch adds support to fsck for per subvolume lost+found.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Dirents currently always point to inodes. Subvolumes add a new type of
dirent, with d_type DT_SUBVOL, that instead points to an entry in the
subvolumes btree, and the subvolume has a pointer to the root inode.
This patch adds bch2_dirent_read_target() to get the inode (and
potentially subvolume) a dirent points to, and changes existing code to
use that instead of reading from d_inum directly.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This patch adds subvolume.c - support for the subvolumes and snapshots
btrees and related data types and on disk data structures. The next
patches will start hooking up this new code to existing code.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Existing quota support breaks badly with snapshots. We're not deleting
the code because some of it will be needed when we reimplement quotas
along the lines of btrfs subvolume quotas.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Figured out the bug we were chasing, and it had nothing to do with
locking btree iterators/paths out of order.
This reverts commit ff08733dd298c969aec7c7828095458f73fd5374.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This patch checks b->hash_val before attempting to lock the node in the
btree, which makes it more equivalent to the "lookup in hash table"
path - and potentially avoids an unnecessary transaction restart if
btree_node_mem_ptr(k) no longer points to the node we want.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Instead of unconditionally upgrading read locks to intent locks in
do_bch2_trans_commit(), this patch changes the path that takes write
locks to first trylock, and then if trylock fails check if we have a
conflicting read lock, and restart the transaction if necessary.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This is a new approach to avoiding the self deadlock we'd get if we
tried to take a write lock on a node while holding a read lock - we
simply upgrade the readers to intent.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
We need to take all needed intent locks when relocking an iterator:
bch2_btree_path_traverse() had a special cased, faster version of this,
but it really should be in up_until_good_node() so that set_pos() can
use it too.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This patch significantly reduces the number of btree lookups required in
the extent update path.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Since btree_path is now internally refcounted, we don't need to clone an
iterator before calling bch2_trans_update() if we'll be mutating it.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
New rule is: if a btree path holds any locks it should be holding
precisely the locks wanted (accoringing to path->level and
path->locks_want).
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Since iter->real_pos was introduced, we no longer have to deal with
extent btree iterators that have skipped past deleted keys - this is a
real performance improvement on btree updates.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
btree_path_traverse_all() traverses btree iterators in sorted order, and
thus shouldn't see transaction restarts due to potential deadlocks - but
sometimes we do. This patch adds some more assertions and tracks some
more state to help track this down.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This splits btree_iter into two components: btree_iter is now the
externally visible componont, and it points to a btree_path which is now
reference counted.
This means we no longer have to clone iterators up front if they might
be mutated - btree_path can be shared by multiple iterators, and cloned
if an iterator would mutate a shared btree_path. This will help us use
iterators more efficiently, as well as slimming down the main long lived
state in btree_trans, and significantly cleans up the logic for iterator
lifetimes.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
If an extent ends up with a replica that is encrypted an a replica that
isn't encrypted (due the user changing options), and then
copygc/rebalance moves one of the replicas by reading from the
unencrypted replica, we had a bug where we wouldn't correctly initialize
op->nonce - for each crc field in an extent, crc.offset + crc.nonce must
be equal.
This patch fixes that by moving op.nonce initialization to
bch2_migrate_write_init.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
When we detect an invalid key being inserted, we should print what code
was doing the update.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
We really only need to distinguish between btree iterators and btree key
cache iterators - this is more prep work for btree_path.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This was used for an optimization that hasn't existing in quite awhile
- iter->uptodate will probably be going away as well.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
These utility functions are for managing btree node state within a
btree_trans - rename them for consistency, and drop some unneeded
arguments.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This is prep work for splitting btree_path out from btree_iter -
btree_path will not have a pointer to btree_trans.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
BTREE_ITER_SET_POS_AFTER_COMMIT is used internally to automagically
advance extent btree iterators on sucessful commit.
But with the upcomnig btree_path patch it's getting more awkward to
support, and it adds overhead to core data structures that's only used
in a few places, and can be easily done by the caller instead.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This consolidates the code for doing extent updates, and makes the btree
iterator usage a bit cleaner and more efficient.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This factors out bch2_dump_trans_iters_updates() from the iter alloc
overflow path, and makes some small improvements to what it prints.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
iter->real_pos needs to match the key returned or bad things will happen
when we go to update the key at that position. When we returned a
pending update from btree_trans_peek_updates(), this wasn't necessarily
the case.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This adds progress stats to sysfs for copygc, rebalance, recovery, and the
cmd_job ioctls.
Signed-off-by: Brett Holman <bholman.devel@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
This fix replaces multiple 64 bit divisions with do_div() equivalents.
Signed-off-by: Brett Holman <bholman.devel@gmail.com>
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
DIV_ROUND_UP() wasn't doing what we wanted when passing it negative
numbers - fix it by just not passing it negative numbers anymore.
Also, no need to do the scaling by compression ratio for incompressible
data.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Valgrind was complaining about a jump depending on uninitialized memory
- we weren't, but this change makes the code less confusing for valgrind
to follow.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
This makes the flow control in bch2_btree_iter_peek() and
bch2_btree_iter_peek_prev() a bit cleaner.
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
