Age | Commit message (Collapse) | Author | |
---|---|---|---|
2018-05-22 | bcachefs: Add another sequence number to struct btree_node | Kent Overstreet | |
Give two different btree nodes for the same part of the keyspace, this lets us tell which one is newer - now, we should have enough information in the btree node header to reconstruct missing interior nodes | |||
2018-05-22 | bcachefs: refactor data move path | Kent Overstreet | |
Prep work for scrubbing | |||
2018-05-22 | bcachefs: Fix allocator striping | Kent Overstreet | |
2018-05-22 | bcachefs: minor prjquota fixes | Kent Overstreet | |
2018-05-22 | bcachefs: -o verbose_init, more error messages | Kent Overstreet | |
for tracing problems with mount | |||
2018-05-22 | bcachefs: fixes for bch2_dev_set_state() | Kent Overstreet | |
2018-05-22 | bcachefs: fix an inconsistency with replicas tracking | Kent Overstreet | |
2018-05-22 | bcachefs: delete dead code | Kent Overstreet | |
replaced by bch2_data_job() | |||
2018-05-22 | bcachefs: BCH_DATA ioctl | Kent Overstreet | |
New infrastructure for jobs that walk existing data and "do stuff" - rereplicate, migrate implemented so far Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> | |||
2018-05-22 | bcachefs: fix fsck error booting in degraded mode | Kent Overstreet | |
also some journal code cleanups | |||
2018-05-22 | bcachefs: fix filesystem character device | Kent Overstreet | |
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> | |||
2018-05-22 | bcachefs: Add superblock field for disk groups | Kent Overstreet | |
2018-05-22 | bcachefs: export bch2_alloc_write() | Kent Overstreet | |
needed for migration tool | |||
2018-05-22 | bcachefs: bch2_getattr() | Kent Overstreet | |
required for proper statx support, apparently | |||
2018-05-22 | bcachefs: don't use bch2_btree_iter_set_pos() unnecessarily | Kent Overstreet | |
it's less efficient | |||
2018-05-22 | bcachefs: call posix_acl_update_mode() from bch2_set_acl | Kent Overstreet | |
fixes xfstest 375 | |||
2018-05-22 | bcachefs: fix a race in bch2_check_mark_super() | Kent Overstreet | |
2018-05-22 | bcachefs: foreground merging of interior btree nodes | Kent Overstreet | |
2018-05-22 | bcachefs: Fix a shutdown bug | Kent Overstreet | |
btree node prefetching can kick off reads in the background - have to flush them | |||
2018-05-22 | bcachefs: More bch2_btree_iter_next optimization | Kent Overstreet | |
gcc apparently doesn't cope well with returning structs by value, even from inline functions | |||
2018-05-22 | bcachefs: more btree_iter refactoring | Kent Overstreet | |
2018-05-22 | bcachefs: optimize bch2_btree_iter_traverse() calls | Kent Overstreet | |
Reduce unnecessary calls, and also call it from bch2_btree_insert_at() when needed Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> | |||
2018-05-22 | bcachefs: btree_iter_next()/btree_iter_next_slot() | Kent Overstreet | |
previously, bch2_btree_iter_advance_pos() would use iter->k to calculate the next iterator position, assuming iter->k had been set by bch2_btree_iter_peek()/peek_with_holes(); this was busted because it was possible in certain codepaths to call bch2_btree_iter_advance_pos() when either iter->k hadn't been initialized, or bch2_btree_iter_traverse() -> btree_iter_down() had scribbled over it. This gets rid of bch2_btree_iter_advance_pos, in favor of bch2_btree_iter_next() and bche_btree_iter_next_slot(), which both just call peek() internall if the iterator isn't uptodate. It's a performance improvement, too. Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> | |||
2018-05-22 | bcachefs: struct btree_iter_level | Kent Overstreet | |
Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com> | |||
2018-05-22 | bcachefs: fix another lockdep splat | Kent Overstreet | |
2018-05-22 | bcachefs: better threshold for compacting whiteouts | Kent Overstreet | |
2018-05-22 | bcachefs: fix a lockdep thing | Kent Overstreet | |
2018-05-22 | bcachefs: fix a faulty assertion | Kent Overstreet | |
2018-05-22 | bcachefs: assorted microoptimizations | Kent Overstreet | |
2018-05-22 | bcachefs: make six locks a bit smaller | Kent Overstreet | |
2018-05-22 | bcachefs: fix another deadlock | Kent Overstreet | |
2018-05-22 | bcachefs: more sysfs debug code | Kent Overstreet | |
2018-05-22 | bcachefs: call journal_reclaim_fast() directly | Kent Overstreet | |
2018-05-22 | bcachefs: better journal_pin_seq() | Kent Overstreet | |
2018-05-22 | bcachefs: fix another deadlock | Kent Overstreet | |
2018-05-22 | bcachefs: fix btree iterators when btree depth == 3 | Kent Overstreet | |
2018-05-22 | bcachefs: fix a deadlock | Kent Overstreet | |
2018-05-22 | bcachefs: disallow POS_MAX keys | Kent Overstreet | |
breaks when we're iterating and call bkey_successor() | |||
2018-05-22 | bcachefs: fix a race in btree_update_drop_new_node() | Kent Overstreet | |
2018-05-22 | bcachefs: provide an obj_cmpfn to rhashtable code | Kent Overstreet | |
default is to use memcmp, which isn't getting inlined (due to kernel not using __builtin_memcpy?) | |||
2018-05-22 | bcachefs: fix shutdown bug(s) | Kent Overstreet | |
2018-05-22 | bcachefs: fix ioctls that set inode flags | Kent Overstreet | |
inode flags that weren't being modified were being cleared | |||
2018-05-22 | bcachefs: fix bch2_bkey_val_to_text() call | Kent Overstreet | |
bch2_bkey_val_to_text() was changed to return number of bytes printed, not buf - but fsck.c calls weren't fixed | |||
2018-05-22 | bcachefs: quota limit enforcement | Kent Overstreet | |
2018-05-22 | bcachefs: don't call page_state_cmpxchg() from bch2_writepage_io_done() | Kent Overstreet | |
it'll make things saner if we can only touch page state under page lock | |||
2018-05-22 | bcachefs: bch2_disk_reservation_init() | Kent Overstreet | |
also fix another bug with reservation nr_replicas being used inconsistently | |||
2018-05-22 | bcachefs: fixes for buffered io & replication | Kent Overstreet | |
2018-05-22 | bcachefs: organize fs-io.c a bit better | Kent Overstreet | |
2018-05-22 | bcachefs: fix build warnings with quotas off | Kent Overstreet | |
2018-05-22 | bcachefs: fix an allocator startup bug | Kent Overstreet | |